Research Question 2#
Which sectors of the economy against which GDP is typically measured, can Air Pollution be the best proxy?
Reviewers: Maxi Leiding, Munir Salihu and Siddharth Saravanan
2.1 Macro-Level Relationship Between NO₂ and GDP#
Key Findings#
[EMRRC23] finds that satellite-based nitrogen dioxide (NO₂) levels strongly correlate with aggregate GDP levels in 178 countries from 2005 to 2020, thereby establishing air pollution as a high frequency indicator of economic activity.
The paper establishes that NO₂ is a better predictor of cyclical economic fluctuations than night-time lights.
Using the COVID-19 lockdowns as an exogenous shock, the study finds that there is a strong negative relationship between the policy response and NO₂ levels. This is because the decrease in NO₂ levels is an instantaneous reflection of the contraction in the economy.
The study’s methodology is based on the observation that NO₂ and night-time lights are collected at different times of the day. Using the technique of identifying the relationship between the two based on the orthogonal measurement errors, the study is able to identify the growth elasticities.
Relevance to This Study#
The evidence establishes that satellite-derived NO₂ is strongly correlated with national GDP levels and short-term economic fluctuations at the aggregate level. While this relationship is identified using country-level data and does not explicitly decompose GDP by sector, the results suggest that NO₂ primarily reflects combustion-intensive components of economic activity. For Research Question 2, this macro-level validation confirms that pollution can serve as a strong proxy for economic output. However, determining which specific GDP sectors drive this relationship requires sectoral analysis, particularly in economies where hydrocarbons, power generation, manufacturing, or transportation dominate production. Therefore, NO₂ is most appropriately interpreted as a proxy for emission-intensive sectors rather than for the entire economy.
2.2 Hydrocarbon Extraction (Oil & Gas Sector)#
Key Findings#
[dGVR+20] shows that TROPOMI provides daily, high-resolution satellite observations of atmospheric methane (CH₄) that clearly detect persistent enhancements over major U.S. oil and gas production regions, such as the Permian Basin.
Methane column increases strongly align with hydrocarbon extraction intensity, demonstrating that satellite-derived CH₄ serves as a sector-specific indicator of oil and gas activity.
Tropospheric NO₂ enhancements are frequently co-located with methane plumes, reflecting combustion-related emissions from drilling equipment, processing operations, and transport infrastructure.
Validation with aircraft and ground-based measurements confirms that the detected methane enhancements originate from boundary-layer emissions, supporting the reliability of satellite-based monitoring for tracking oil and gas sector emissions.
Relevance to This Study#
Methane (CH₄) is shown to be a very specific indicator of oil and gas extraction, rather than a general measure of industrial or transportation activity. Methane levels increase closely with upstream oil and gas production, which makes CH₄ a useful satellite-based proxy for the extractive sector’s contribution to GDP. This is especially important for countries like Algeria, where hydrocarbons make up a large share of economic output. In countries such as Myanmar, where official data and monitoring systems are limited, satellite-based methane data can help identify oil and gas activity more reliably.
2.3 Power Generation & Heavy Industry#
Key Findings#
[FMK+23] develops an updated global catalogue of major sulfur dioxide (SO₂) emission sources derived entirely from satellite observations covering the period 2005–2021.
The catalogue identifies 759 persistent SO₂ point sources worldwide and classifies them into volcanoes, coal- and oil-fired power plants, metal smelters, and oil and gas facilities.
Emissions for each source are quantified using physically based plume-fitting techniques that combine satellite-derived SO₂ column densities with ERA5 wind field data, enabling estimation of source-specific emission rates.
The study documents an approximate 50% decline in global anthropogenic SO₂ emissions since the mid-2000s, largely attributable to regulatory interventions in power generation and heavy industry.
Relevance to This Study#
This study provides the technical foundation for assigning pollution to specific GDP sectors. It shows that SO₂ is strongly linked to electricity generation, metal smelting, and hydrocarbon processing, rather than to transportation or residential activity. Because these sources are large and fixed, satellite data can clearly identify them. For Algeria, SO₂ point-source data helps distinguish emissions from power plants and hydrocarbon facilities, strengthening sectoral GDP analysis beyond what NO₂ alone can provide. In Myanmar, the global catalogue supports the identification of industrial activity at the sub-national level, particularly where official data and monitoring systems are limited.
2.4 Manufacturing, Mining & Industrial Production#
Key Findings#
[BMP21] study demonstrates a strong and statistically significant relationship between satellite-derived tropospheric NO₂ concentrations and monthly industrial production across both advanced and emerging economies, showing that atmospheric pollution can serve as a real-time indicator of economic activity.
NO₂ is identified as a particularly informative proxy for manufacturing because it is closely linked to fossil fuel combustion in factories, power plants, and industrial supply chains, making it sensitive to fluctuations in production intensity.
Incorporating high-frequency (daily) NO₂ satellite data significantly improves short-term industrial production nowcasts compared to traditional autoregressive models and commonly used survey indicators such as the Purchasing Managers’ Index (PMI).
The authors apply machine learning methods, including k-nearest neighbors for spatial interpolation and random forests for meteorological normalization, to isolate the economic signal from weather-related variation in pollution data.
The elasticity between NO₂ and industrial production is stronger in countries with a larger manufacturing share of GDP, confirming that pollution-based indicators are most informative where industry plays a dominant economic role.
[HWDW22] shows that manufacturing agglomeration significantly increases air pollution, while the agglomeration of producer services (such as finance and IT) reduces pollution levels, highlighting a structural difference between industrial and service sectors.
The study finds that pollution levels are positively associated with heavy industrial activity, including steel, cement, and energy production, but weakly or negatively correlated with service-sector expansion.
These findings suggest that NO₂ is more closely linked to manufacturing-based GDP components than to service-sector output.
[PN25] provides direct evidence of the elasticity between NO₂ concentrations and sectoral GDP, showing particularly strong relationships for mining, utilities, and manufacturing.
The study demonstrates that the effectiveness of NO₂ as a GDP proxy depends on spatial scale, with coarser regional data producing stronger and more stable relationships than very high-resolution data.
The results confirm that NO₂ is most reliable as a proxy for combustion-intensive sectors of the economy rather than for service-based or digital economic activity.
Relevance to This Study#
The literature shows that nitrogen dioxide (NO₂) is closely linked to industrial and manufacturing activity, but it also highlights important limitations. Pollution levels increase with heavy industrial production, mining, and utilities, while service-sector expansion may reduce or weaken the pollution signal. This means that falling NO₂ does not necessarily indicate economic decline; it may instead reflect structural shifts toward less emission-intensive sectors. Therefore, pollution data must be interpreted carefully, especially in economies undergoing industrial transformation.
At the same time, empirical evidence confirms that NO₂ is a strong proxy for industrial GDP components and often performs better than nighttime lights in capturing combustion-based economic activity. However, its effectiveness depends on spatial scale, with coarser regional data generally producing more stable results than very high-resolution data.In Algeria, NO₂ levels are likely to move with changes in industrial and energy activity because these sectors rely heavily on fuel combustion. In Myanmar, where detailed and frequent industrial data are limited, satellite-based NO₂ can help track economic activity. However, using NO₂ together with other pollutants can give a clearer picture of which sectors are driving changes in the economy.
2.5 Construction & Infrastructure Development#
Key Findings#
[Far25] isolates Nitrogen Dioxide (NO₂) as the primary pollutant associated with construction activity, finding no significant causal link with SO₂ or PM.
The study further demonstrates that this strong correlation is driven by the engineering characteristics of heavy-duty diesel construction equipment, such as excavators, bulldozers, and cranes. These machines operate at high load factors and have disproportionately higher NOₓ emission factors relative to PM₂.₅, making NO₂ a more precise indicator of active infrastructure development compared to particulate pollutants.
Relevance to This Study#
The literature shows a strong link between construction activity and increases in nitrogen dioxide (NO₂) levels. Heavy-duty diesel equipment used in construction, such as excavators and bulldozers, produces high NOₓ emissions, leading to a noticeable rise in NO₂ when construction activity increases. This increase is stronger than what is observed for other pollutants like PM or SO₂. This means construction-sector GDP can be compared with NO₂ levels to see whether pollution reflects growth in physical investment, especially in developing countries where detailed construction data may be limited.
Summary of Key Findings#
The literature shows that air pollution serves as a reliable proxy for economic activity primarily in combustion-intensive sectors of GDP. Satellite-observed nitrogen dioxide (NO₂) is strongly associated with manufacturing, industrial production, mining, utilities, transportation, and construction, as these sectors rely heavily on fossil-fuel combustion and energy-intensive processes. Methane (CH₄) is particularly effective in tracking hydrocarbon extraction, while sulfur dioxide (SO₂) provides a clear signal of power generation and certain heavy industrial operations.
At the macro level, NO₂ closely follows aggregate GDP fluctuations in economies where industrial and energy sectors account for a substantial share of output. In contrast, pollution is a much weaker indicator for service-based sectors such as finance, information technology, and public administration, which generate relatively low direct emissions. At the same time, pollution can negatively affect productivity, especially in labor and cognition intensive service sectors. Taken together, the evidence suggests that air pollution is most effective as a sector-specific proxy for pollution-intensive economic activity, and its usefulness depends on the underlying structure of the economy.
References#
Ezran, I., Morris, S. D., Rama, M., & Riera-Crichton, D. (2023). Measuring global economic activity using air pollution (Policy Research Working Paper No. 10445). World Bank. https://doi.org/10.1596/1813-9450-10445
Bricongne, J.-C., Meunier, B., & Pical, T. (2021). Can satellite data on air pollution predict industrial production? Banque de France Working Paper No. 847. https://publications.banque-france.fr/en/can-satellite-data-air-pollution-predict-industrial-production
de Gouw, J. A., Veefkind, J. P., Roosenbrand, E., Dix, B., Lin, J. C., Landgraf, J., & Levelt, P. F. (2020). Daily satellite observations of methane from oil and gas production regions in the United States. Scientific Reports, 10(1), Article 1379. https://doi.org/10.1038/s41598-020-57678-4
Fioletov, V. E., McLinden, C. A., Krotkov, N. A., et al. (2023). Version 2 of the global catalogue of large anthropogenic and volcanic SO₂ sources and emissions derived from satellite measurements. Earth System Science Data, 15, 75–103. https://essd.copernicus.org/articles/15/75/2023/
He, W., Wang, B., Danish, & Wang, Z. (2022). Impact of collaborative agglomeration of manufacturing and producer services on air quality: Evidence from the Yangtze River Economic Belt. Atmosphere, 13(6), 966. https://doi.org/10.3390/atmos13060966
Parubets, S., & Naito, H. (2025). Predicting economic activity using atmospheric nitrogen dioxide (NO₂) satellite data: Evidence from local economic indicators in Japan. PLOS One, 20(12), e0337901. https://doi.org/10.1371/journal.pone.0337901
Farooqui, M. A. (2025). Construction sector and air pollution: Evidence from India. Ideas for India. https://www.ideasforindia.in/