Research

Urban air pollution hotspots pose significant health risks, yet their detection and analysis remain limited by the sparsity of public sensor networks. This paper addresses this challenge by combining predictive modeling and mechanistic approaches to comprehensively monitor pollution hotspots. We enhanced New Delhi’s existing sensor network with 28 low-cost sensors, collecting PM2.5 data over 30 months from May 1, 2018, to Nov 1, 2020. Applying established definitions of hotspots to this data, we found the existence of additional 189 hidden monthly hotspots in addition to confirming the 660 detected by the public network. Using predictive techniques like Space-Time Kriging, we identified monthly hotspots with 95% precision and 88% recall with 50% sensor failure rate, and with 98% precision and 95% recall with 50% missing sensors. The projections of our predictive model were further compiled into policy recommendations for public authorities. Additionally, we developed a Gaussian Plume Dispersion Model to understand the mechanistic underpinnings of hotspot formation, incorporating an emissions inventory derived from local sources. Our mechanistic model is able to explain 65% of observed transient hotspots. Our findings underscore the importance of integrating data-driven predictive models with physics-based mechanistic models for scalable and robust air pollution management in resource-constrained settings.

Market-based environmental regulations are seldom used in low-income countries, where pollution is highest but state capacity is often low. We collaborated with the Gujarat Pollution Control Board (GPCB) to design and experimentally evaluate the world’s first particulate-matter emissions market, which covered industrial plants in a large Indian city. There are three main findings. First, the market functioned well. Treatment plants, randomly assigned to the emissions market, traded permits to become significant net sellers or buyers. After trading, treatment plants held enough permits to cover their emissions 99% of the time, compared with just 66% compliance with standards under the command-and-control status quo. Second, treatment plants reduced pollution emissions, relative to control plants, by 20%–30%. Third, the market reduced abatement costs by an estimated 11%, holding constant emissions. This cost-savings estimate is based on plant-specific marginal cost curves that we estimate from the universe of bids to buy and sell permits in the market. The combination of pollution reductions and low costs imply that the emissions market has mortality benefits that exceed its costs by at least 25 times.

We study carbon offsets sold by firms in China under the Clean Development Mechanism (CDM). We find that offset-selling firms, meant to cut carbon emissions, instead increase them by 49% after starting an offset project. In a model of firm investment decisions and offset review, we estimate that CDM firms increase emissions due to both the selection of higher-growth firms into projects (35 pp) and because offset projects themselves boost firm growth and therefore emissions (14 pp). The CDM reduces global surplus by causing damages from increased emissions four times greater than private gains from trade in the offset market.

Climate policies vary widely across countries, with some countries imposing stringent emissions policies and others doing very little. When climate policies vary across countries, energy-intensive industries have an incentive to relocate to places with few or no emissions restrictions, an effect known as leakage. Relocated industries would continue to pollute but would be operating in a less desirable location. We consider solutions to the leakage problem in a simple setting where one region of the world imposes a climate policy and the rest of the world is passive. We solve the model analytically and also calibrate and simulate the model. Our model and analysis imply: (1) optimal climate policies tax both the supply of fossil fuels and the demand for fossil fuels; (2) on the demand side, absent administrative costs, optimal policies would tax both the use of fossil fuels in domestic production and the domestic consumption of goods created with fossil fuels, but with the tax rate on production lower due to leakage; (3) taxing only production (on the demand side), however, would be substantially simpler and almost as effective as taxing both production and consumption, because it would avoid the need for border adjustments on imports of goods; and (4) the effectiveness of the latter strategy depends on a low foreign elasticity of energy supply, which means that forming a taxing coalition to ensure a low foreign elasticity of energy supply can act as a substitute for border adjustments on goods.

The use of air quality monitoring networks to inform urban policies is critical especially where urban populations are exposed to unprecedented levels of air pollution. High costs, however, limit city governments’ ability to deploy reference grade air quality monitors at scale; for instance, only 33 reference grade monitors are available for the entire territory of Delhi, India, spanning 1500 sq km with 15 million residents. In this paper, we describe a high-precision spatio-temporal prediction model that can be used to derive fine-grained pollution maps. We utilize two years of data from a low-cost monitoring network of 28 custom-designed low-cost portable air quality sensors covering a dense region of Delhi. The model uses a combination of message-passing recurrent neural networks combined with conventional spatio-temporal geostatistics models to achieve high predictive accuracy in the face of high data variability and intermittent data availability from low-cost sensors (due to sensor faults, network, and power issues). Using data from reference grade monitors for validation, our spatio-temporal pollution model can make predictions within 1-hour time-windows at 9.4, 10.5, and 9.6% Mean Absolute Percentage Error (MAPE) over our low-cost monitors, reference grade monitors, and the combined monitoring network respectively. These accurate fine-grained pollution sensing maps provide a way forward to build citizen-driven low-cost monitoring systems that detect hazardous urban air quality at fine-grained granularities.

The economic impacts of climate change are highly uncertain. Two of the most important uncertainties are the sensitivity of the climate system and the so-called damage functions, which relate climate change to economic costs and benefits. Despite broad awareness of these uncertainties, it is unclear which of them is most important, especially at the regional level. Here we construct regional damage functions, based on two different global damage functions, and apply them to two climate models with vastly different climate sensitivities. We find that uncertainty in both climate sensitivity and aggregate economic damages per degree of warming are of similar importance for the global economic impact of climate change, with the decrease in global economic productivity ranging between 4% and 24% by the end of the century under a high-emission scenario. At the regional level, however, the effects of climate change can vary even more substantially, depending both on a region's initial temperature and the amount of warming it experiences, with some regions gaining in productivity and others losing. The ranges of uncertainty are therefore potentially much larger at a regional level. For example, at the end of the century, under a high-emission scenario, we find that India's productivity decreases between 13% and 57% and Russia's increases between 24% and 74%, while Germany's change in productivity ranges from an increase of 8% to a decrease of 4%. Our findings emphasize the importance of including these uncertainties in estimates of future economic impacts, as they are vital for the resulting impacts and thus policy implications.

Common resources may be managed with inefficient policies for the sake of equity. We study how rationing the commons shapes the efficiency and equity of resource use in the context of agricultural groundwater use in Rajasthan, India. We find that rationing binds on input use, such that farmers, despite trivial prices for water extraction, use roughly the socially optimal amount of water on average. The rationing regime is still grossly inefficient, because it misallocates water across farmers, lowering productivity. Pigouvian reform would increase agricultural surplus by 12% of household income yet fall well short of a Pareto improvement over rationing.

Can targeting information to network-central farmers induce more adoption of a new agricultural technology? By combining social network data and a field experiment in 200 villages in Malawi, we find that targeting central farmers is important to spur the diffusion process. We also provide evidence of one explanation for why centrality matters: a diffusion process governed by complex contagion. Our results are consistent with a model in which many farmers need to learn from multiple people before they adopt themselves. This means that without proper targeting of information, the diffusion process can stall and technology adoption remains perpetually low.

The integration of markets may improve efficiency by lowering costs or reducing local market power. India, seeking to reduce electricity shortages, set up a new power market, in which transmission constraints sharply limit trade between regions. During congested hours, measures of market competitiveness fall and firms raise bid prices. I use confidential bidding data to estimate the costs of power supply and simulate market outcomes with more transmission capacity. Counterfactual simulations show that transmission expansion increases market surplus by 22 percent, enough to justify the investment. One-third of this gain is due to sellers' response to a more integrated grid.