CLEARING THE AIR: A TALE OF THREE CITIES | a b | CLEARING THE AIR: A TALE OF THREE CITIES ©2020 The World Bank Group 1818 H Street NW Washington, DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org All rights reserved. This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions: The material in this work is subject to copyright Figures B2.1, B2.2, and 11 have been reproduced with permission from the authors. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. Attribution: World Bank. 2020. Clearing the Air: A tale of three cities. ©World Bank Editor: Simi Mishra Designed and Printed: Roots Advertising Photo Credit: [Cover photo: ESB Professional/Shutterstock.com, Azhar_khan/Shutterstock.com, Gill_figueroa/Shutterstock.com] Marianna Ianovska/Shutterstock.com, Rudra Narayan Mitra/Shutterstock.com, Daniel Prudek/Shutterstock. com, testing/ Shutterstock.com, pinholeimaging/Shutterstock.com, 4H4 Photography/Shutterstock.com, Kamira/Shutterstock.com, HelloRF Zcool/Shutterstock.com, AJP/Shutterstock.com CLEARING THE AIR A tale of three cities Contents Acknowledgements v Abbreviations and Acronyms vii EXECUTIVE SUMMARY 1 1. AIR POLLUTION: A GLOBAL CHALLENGE 7 2. WHAT IS THE RELATIONSHIP BETWEEN AIR QUALITY AND 14 ECONOMIC GROWTH? 3. TACKLING AIR POLLUTION: LESSONS FROM MEXICO CITY, 22 BEIJING, AND DELHI 3a. Air Quality Management in Mexico City 23 3b. Air Quality Management in Beijing and the Greater Jing-Jin-Ji (JJJ) Region 30 3c. Air Quality Management in Delhi 37 4. WHAT LESSONS CAN OTHER COUNTRIES DRAW? 47 REFERENCES 52 ANNEX: POLLUTION INTENSITY OF ECONOMIC GROWTH IN SOUTH 56 ASIA AND OTHER REGIONS BOXES BOX 1: Air Quality Trends in India 10 BOX 2: Explaining two paradoxes about pollution intensity of India’s growth path 18 BOX 3: Controlling Air Pollution from Coal-Fired Power in India 38 BOX 4: Pradhan Mantri Ujjwala Yojana Expands Clean Cooking in India 45 FIGURES FIGURE ES1: Air pollution is a challenge across the world 1 FIGURE ES2: Air pollution is the fourth largest health risk globally 2 FIGURE ES3: India is on a pollution-intensive growth path, driven by states in the 3 Indo-Gangetic FIGURE ES4: High pollution intensive growth path is not the norm 3 FIGURE ES5: Key Components of Air Quality Management 4 FIGURE 1: Mean annual ambient PM2.5 pollution across the world, 1990-2015 9 FIGURE B1.1: Monitored PM2.5 pollution in 47 Indian cities, January-November 2018 10 FIGURE B1.2: Mean annual ambient PM2.5 pollution across the Indo-Gangetic Plain, 11 1990-2015 FIGURE 2: Leading fatal health risks globally, 2017 12 FIGURE 3: Mean annual ambient PM2.5 pollution versus GNI per capita, 15 1990 and 2015 FIGURE 4: GDP per capita versus mean annual ambient PM2.5 in select large 16 middle-income countries, 1990-2015 FIGURE 5: GDP per capita versus mean annual ambient PM2.5 in select large 17 middle-income countries, 1990-2015 FIGURE B2.1: State-wise origin of mean annual population-weighted ambient 18 PM2.5 by sector, 2015 FIGURE B2.2: Contribution of local, regional, and transboundary sources of 19 ambient PM2.5 exposure in Indian states, 2015 FIGURE 6: Mean annual PM2.5 in 2015 (left) and change in mean annual PM2.5 20 from 1990 to 2015 (right) for low- and middle-income countries with average annual GDP per capita growth of at least 3 percent FIGURE 7: Monitored concentrations of pollutants in Mexico City compared 24-25 to national standards and WHO guidelines, January 1986 to September 2018 FIGURE 8: Monthly PM2.5 concentrations in Beijing, February 2009 to October 2018 30 FIGURE 9: Timeline of air quality management actions in Beijing, 1998-2013 31 FIGURE 10: Daily average PM2.5 concentrations during severe pollution 32 episodes in Beijing (January 2013) and New Delhi (November 2017) FIGURE 11: Sources of PM2.5 pollution in cities in the JJJ region 34 FIGURE B3.1: Share of coal-fired power in total electricity generation in India, 38 1980-2018 FIGURE 12: Monitored concentrations of pollutants in Delhi compared to national 42-43 standards, 1990-2018 FIGURE 13: Sources of PM2.5 pollution in Delhi NCR region, 2017-18 44 FIGURE 14: Key Components of Air Quality Strategy 48 TABLES TABLE 1: Mean annual population-weighted exposure to ambient PM2.5 per region, 9 1990-2015 TABLE 2: Mean annual population-weighted exposure to ambient PM2.5 per 15 income group and for China, India, and Mexico, 1990-2015 TABLE 3: Primary Emissions from Mobile Sources in the Mexico City 26 Metropolitan Area TABLE 4: Timing of Mexico’s Emission Standards and Alignment with 27 Standards from the United States and the European Union TABLE 5: Driving Restrictions in Mexico City 28 TABLE B3.1: Power Plant Emissions Standards in Various Countries 39 TABLE 6: Timeline of key measures to tackle air pollution in Delhi in the 1990s and early 2000s TABLE A.1: Generalized spatial two-stage least squares (GS2SLS) regression results 59 TABLE A.2: Maximum likelihood (ML) estimator regression results 60 TABLE A.3: Estimated elasticity values of mean annual PM2.5 with respect to GDP 61 per capita for countries at India’s income level in 1995 (a) versus 2015 (b), in and outside the South Asia region Acknowledgements This report was prepared by a team led by Urvashi Narain comments were also received from participants at the and composed of Christopher Sall, Jostein Nygard, meetings held by the Department of Economic Affairs, Dafei Huang, Ernesto Sanchez Triana, and Katharina Ministry of Finance, Government of India on March 15, Siegmann. Contributions were also received from Pedro 2019 and July 15, 2019 to discuss the draft report, and in Arizti, Sharlene Chichgar, Momoe Kanada, Heey Jin Kim writing from Ministry of Agriculture & Farmers Welfare, and Ishaa Srivastava. Ministry of Coal, Ministry of Health & Family Welfare, Ministry of Heavy Industries & Public Enterprises, The team is very grateful for the support and overall Ministry of Mines, Ministry of Petroleum & Natural guidance received from Karin Kemper (Global Director, Gas, Ministry of Power, Ministry of Road Transport & World Bank), Junaid Ahmad (Country Director, World Highways, Niti Aayog, Climate Change Finance Unit, Bank), John Roome (Regional Director, World Bank), and Office of Indian Executive Director, World Bank. Christophe Crepin (Practice Manager, World Bank), Comments on the Mexico case study were provided by Magda Lovei (Practice Manager, World Bank), and Kseniya Eduardo Olivares Lechuga (National Institute of Ecology Lvovsky (Practice Manager, World Bank). Constructive and Climate Change), Victor Paramo Figueroa and comments on the report were received from the following Ramiro Barrios Castrejon (Secretary of Environment and peer reviewers: Helena Naber, Garo Batmanian, Gailius Natural Resources -SEMARNAT), the Environmental Draugelis, Michael Toman, and Madhur Gautam. The Commission of the Megalopolis (CAMe), and Santiago team would also like to acknowledge the suggestions Enriquez and Mariana Aguirre (World Bank) for which received from several other colleagues, including the team is grateful. Special thanks to Nitika Man Poonam Gupta, Aurelien Kruse, Sumila Gulyani, Sudip Singh Mehta and Latha Sridhar for their support in the Mazumdar, Rinku Murgai, Charles Undeland, Luc publication of the report. Lecuit, and Andrew Zakharenka. Ajay Mathur (Director General, The Energy Research Institute), Anumita Roy Any remaining errors or omissions are the authors’ own. Chowdhury (Executive Director, Centre for Science and Environment), and Mukesh Kumar (Professor, The team also recognizes the financial support received Indian Institute of Technology, Kanpur) also provided from the World Bank’s Pollution Management and comments on an earlier draft. Extensive constructive Environmental Health Trust Fund. Abbreviations and Acronyms CAA Clean Air Act CAAQM Continuous Ambient Air Quality Monitoring CAM Metropolitan Environmental Commission CEEW Council on Energy, Environment and Water CEM Continuous Emissions Monitoring CNG Compressed Natural Gas CPCB Central Pollution Control Board EKC Environmental Kuznets Curve EPA United States Environmental Protection Agency EPCA Environmental Pollution (Prevention and Control) Authority GDP Gross Domestic Product IGP Indo-Gangetic Plain IIASA International Institute for Applied Systems Analysis JJJ Jing-Jin-Ji LPG Liquefied Petroleum Gas MCMA Mexico City Metropolitan Area MoEFCC Ministry of Environment, Forest and Climate Change NAAQS National Ambient Air Quality Standards NAMP National Air Quality Monitoring Programme NCR National Capital Region PICCA Integrated Program against Atmospheric Pollution in the Mexico City Metropolitan Area PPP Purchasing Power Parity PROAIRE Program to Improve Air Quality SC Supreme Court SIP State Implementation Plan SPCB State Pollution Control Board UT Union Territory WHO World Health Organization WTP Willingness to pay All dollar amounts are in US dollars, unless otherwise indicated. viii | CLEARING THE AIR: A TALE OF THREE CITIES Executive Summary Air pollution is a major health risk, and a drag on a small particulates with a diameter of less than 2.5 country’s development. microns, about one-thirtieth the width of a human hair. Exposure to PM2.5 can cause such deadly illnesses as Air pollution presents an increasingly apparent challenge lung cancer, stroke, and heart disease, and the WHO has to health and development across the globe (see Figure recommended that people should not be exposed ES1). In 2015, the latest year for which global coverage to concentrations of PM2.5 pollution higher than 10 of air quality is available, about 94 percent of the world’s micrograms per cubic meter (µg/m3) on average each people resided in areas for which the average annual year (WHO, 2005). In 2015, the mean annual exposure PM2.5 exceeded the World Health Organization’s (WHO) for countries in South Asia, and in the Middle East and guideline value. This challenge is only growing in a North Africa was 77 µg/m3, almost eight times the WHO number of low and lower-middle income countries. Air guideline values. quality has deteriorated across many of these countries since the 1990s, and their population is being exposed Exposure to PM2.5 is a major health risk. Worldwide, an to increasing and unhealthy levels of ambient PM2.5, estimated 4.13-5.39 million people died prematurely in FIGURE ES1: Air pollution is a challenge across the world (Mean annual ambient PM2.5 pollution across the world, 1990-2015) Mean annual PM2.5 in 1990 Mean annual PM2.5 in 2015 Change from 1990 to 2015 Mean annual PM2.5 (top) (micrograms per cubic meter) 0-10 (WHO guideline) 50 - 75 10 - 35 (WHO IT-1) 75 - 100 35 - 50 Above 100 -15 - 0 (decrease) 10 - 20 0-5 20+ (increase) Sources: IHME (2017), van Donkelaar et al. (2016), 5 -10 Shaddick et al. (2018) 2 | CLEARING THE AIR: A TALE OF THREE CITIES 2017 from exposure to PM2.5 pollution. About 8 percent Countries appear to follow growth paths with of all attributable deaths globally in 2017 were thus linked different levels of pollution intensity, suggesting to PM2.5 pollution (Figure ES2), more than the number that policy decisions, investments, and technologies of people who died from HIV/AIDS, tuberculosis, and all have an important role to play in affecting the malaria combined.1 pollution intensity of growth, and that countries cannot simply grow their way out of pollution. The health impacts of pollution also represent a heavy cost to the economy. Lost labor income due to fatal illness Figure ES3 (left panel) illustrates the relationship between from PM2.5 pollution globally in 2017 was in the range of mean annual PM2.5 exposure and the level of income, as US$ 131-317 billion,2 equal in magnitude to about 0.1- measured by GDP per capita, for large middle-income 0.3 percent of GDP. Beyond reduced labor earnings, when countries from 1990 to 2015. the broader costs of fatal illness to people’s wellbeing are measured—following a method adopted by public Pollution in the countries shown in the lower part of the agencies in many countries—the damages from PM2.5 figure appears to have already reached a turning point. pollution are equal in magnitude to 1.9 percent of On the other hand, as seen in the upper left of the figure, GDP. Air pollution is also likely reducing agricultural India, Nepal, and Pakistan appear to be on an entirely productivity. One study found that ozone and black different, more pollution-intensive path, with no obvious carbon (emitted mostly from household cookstoves) turning point. More granular state-level analysis reveals cut yields of wheat and rice by about 33 percent and 22 that India’s pollution-intensive growth pattern is driven percent, respectively, in India’s largest producing states in primarily by trends in the Indo-Gangetic Plain (IGP), 2010. Lower yields have translated into an annual loss of including in Bihar, Delhi, Haryana, Jharkhand, Punjab, 24 million tons of harvested wheat alone, worth about Uttar Pradesh, and West Bengal. These seven states had US$ 5 billion (Burney and Ramanathan, 2015). the highest elasticities of PM2.5 with respect to income FIGURE ES2: Air pollution is the fourth largest health risk globally (All cause mortality globally in 2017) Metabolic risks 31.4 Dietary risks 19.5 Tobacco 14.5 Air pollution (PM2.5) 8.2 Child and maternal malnutrition 5.7 WASH 2.9 Low physical activity 2.3 0 5 10 15 20 25 30 35 Share of all-cause mortality globally in 2017(%) Notes: “Air pollution (PM2.5)” includes ambient PM2.5 pollution and household PM2.5 pollution from cooking with solid fuels; “WASH” includes unsafe water, sanitation, and handwashing. Source: Institute for Health Metrics and Evaluation, Global Burden of Disease Study 2017 (2018); Stanaway et al. (2018). 1 The range represents the 95-percent uncertainty interval. Estimates are from the Global Burden of Disease Study 2017 (GBD 2017) and include exposure to ambient PM2.5 pollution as well as PM2.5 in households cooking with solid fuels (Stanaway et al., 2018). 2 The estimates of the costs of pollution in terms of lost labor income and welfare loss are based on the methodology used in World Bank-IHME (2016) and detailed in Narain and Sall (2016). Damages are expressed in constant US dollars at purchasing power parity (PPP) and year 2011 prices. Both the costs of fatal illness from ambient PM2.5 and household PM2.5 from cooking with solid fuels are included. CLEARING THE AIR: A TALE OF THREE CITIES | 3 FIGURE ES3: India is on a pollution-intensive growth path, driven by states in the Indo-Gangetic Plain (GDP per capita versus mean annual ambient PM2.5 in select large middle-income countries, 1990-2015) 80 160 Nepal PakistanIndia Delhi 70 140 Mean annual PM2.5,1990-2015 (μg/m3) 60 120 Bihar China Uttar Pradesh Mean annual PM2.5 (μg/m3) Haryana 50 Iran, Islamic Rep. 100 Myanmar West Bengal Punjab Jharkhand 40 80 South Africa Turkey 30 Sri Lanka 60 China Vietnam Thailand 20 Ukraine Mexico Kazakhstan 40 10 Brazil 20 Mexico 0 0 0 5,000 10,000 15,000 20,000 25,000 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 GDP per capita, 1990-2015 (year 2011 US$, PPP) GDP per capita (2011 US$, PPP) Source: Global Burden of Disease Study 2016 data provided by Institute for Health Metrics and Evaluation (IHME); estimates by IHME (2017), van Donkelaar et al. (2016), and Shaddick et al. (2018); GDP per capita data from World Bank, World Development Indicators database. FIGURE ES4: High pollution intensive growth path is not the norm (Change in mean annual PM2.5 from 1990 to 2015 for low- and middle-income countries with average annual GDP per capita growth of at least 3 percent) 40 Bangladesh Nigeria Change in mean annual PM2.5, 1990-2015 (μg/m3) 30 Chad 20 Iraq Sudan India Ethiopia Uganda 10 Mongolia Armenia China Malaysia Lao PDR Azerbaijan Myanmar Cambodia 0 Vietnam Thailand Albania Uzbekistan Sri Lanka Bosnia and Serbia -10 Herzegovina Romania Belarus Moldova -20 Bulgaria -30 2 3 4 5 6 7 8 9 10 Average GDP per capita growth rate, 1990-2015 (%) Source: Global Burden of Disease Study 2016 data provided by Institute for Health Metrics and Evaluation (IHME); estimates by IHME (2017), van Donkelaar et al. (2016), and Shaddick et al. (2018); GDP per capita data from World Bank, World Development Indicators database. 4 | CLEARING THE AIR: A TALE OF THREE CITIES (that is, the largest increases in pollution per unit increase FIGURE ES5: Key Components of Air Quality in income)—although at markedly different levels of Management per capita income. Bihar and Jharkhand had the lowest levels of GDP per capita in the country, while Delhi and Haryana had the highest (see right panel of Figure ES3). The high pollution intensity of economic growth for Information South Asian countries is not the norm. Many low- and middle-income countries that have experienced rapid income growth have achieved a reduction in ambient PM2.5. Only six of the 42 low- and middle-income Air Quality countries with average annual rates of GDP per capita Management growth higher than 3 percent between 1990 and 2015 saw air quality deteriorate as much as it did in India. All the Incentives Institutions countries with GDP per capita growth rates higher than India saw smaller increases or decreases in ambient PM2.5 (see Figure ES4). These trends suggest that policy decisions, investments, and technologies have a role to play in bending, flattening program (Bell et al. 2004). The India National Air Quality or shifting the Environmental Kuznets Curve – the notion Index program, initiated in 2015, is an important step that pollution first worsens and then improves at higher towards supporting action. Additionally, apart from levels of income as a country develops. The experiences better monitoring data, it is important to improve of three cities – Mexico City, Beijing, and Delhi – offers data to inform air quality action plans, which, in turn, some lessons on how countries can tackle the growing require data on pollution sources and cost effectiveness challenge of air pollution. of different policy interventions. Data on emissions and sources of pollution allow policymakers to build models There is no silver bullet, and air pollution will only to assess expected improvements from current and be tackled through sustained political commitment. planned policy interventions, to set measurable targets Information, incentives, and institutions are the three and to identify strategies to meet targets in a cost-effective prongs of an effective air pollution management manner, and monitor progress. Finally, timely and strategy for any country (see Figure ES5). accessible data can support enforcement of regulations. Installing emissions monitors in large industrial facilities Information: adequate and accessible and power plants and making these data public can help hold local regulators and plant operators accountable for Data on air pollution concentrations and its health upholding environmental standards, as has been the case implications, on sources of pollution, on violations in China. India’s policy requiring polluting industries of regulations, etc. are critical to the design and to install Online Continuous Emissions/Effluent implementation of air quality programs. Expanding Monitoring Systems will similarly improve enforcement air quality monitoring networks, supporting public of existing regulations. disclosure of information on air quality levels, and raising awareness on the health and other economic Incentives: mainstreamed costs of air pollution have been found to increase demand for action. In Mexico City, for example, careful Countries need a strong regulatory mechanism to ensure analysis of the impacts of air pollution on the health of that states and cities are incentivized to implement policies children helped galvanize public support for the city’s and programs to reduce air pollution. Be it carrot- or stick- first air quality management strategy. In Delhi too, based, a mechanism to incentivize implementation of air overall public support and awareness about the health quality management plans is needed. An examination impacts of air pollution contributed to the government of the role of the government and the Supreme Court throwing its support behind the CNG conversion in the efforts to clean Delhi’s air points to a pattern CLEARING THE AIR: A TALE OF THREE CITIES | 5 of dependence on the courts for compliance. Time central government provided US$ 9.29 billion in special and time again, the government announced measures funds and budgetary resources to support air quality to reduce pollution but did not follow through on management in the region, including Beijing and implementation. The Supreme Court then weighed in to the surrounding provinces and cities. These financial force the government to implement the policy measures resources were used to support a variety of incentive it had previously announced. Should the Supreme programs, including subsidies for end-of-pipe controls Court continue to play this role, or can a stronger and boiler retrofits in power plants and factories, rebates legal framework provide a mechanism to incentivize for scrapping older vehicles, and payments to households governments to implement policies designed to tackle switching out coal-fired heating stoves for gas or electric air pollution? Sanction powers granted to the United systems. Provinces, moreover, committed their own States Environment Protection Agency (EPA) under the resources and used their own and centrally-allocated United States’ Clean Air Act (CAA) offer some lessons funds to leverage additional financing from the private to countries on how to incentivize implementation. As sector to the order of US$ 2.96 billion. In the mid-2000s, per the provisions of the CAA, if an area or city is found Mexico City provided direct subsidies to drivers of old to have pollution levels above acceptable standards, taxis in exchange for retiring and scrapping their old they are required to prepare and submit to the EPA a vehicles, along with access to low-cost loans for vehicle State Implementation Plan (SIP). The SIP provides a renovations or purchase of more efficient vehicles. time bound set of measures, potentially to be imposed Similarly, a range of incentives were offered to encourage on industry, transportation, etc., that are necessary to industrial enterprises to make the switch from fuel oil to achieve compliance with air quality standards. The CAA, natural gas and to install emissions control equipment. however, includes additional provisions to enforce SIP Fiscal incentives and exemptions from emergency implementation. Namely, if the state fails to submit an restrictions were included which require industrial acceptable plan or fails to implement the measures of an plants to curtail their production when air pollution approved plan, the CAA empowers EPA to impose one reaches high levels. of two sanctions: (i) withholding certain federal highway funds by prohibiting the Secretary of Transportation from Institutions: fit-for-purpose awarding funds from the Federal-aid Highway Program; The multi-jurisdictional nature of air pollution requires or (ii) imposing a “2:1 offset” requirement on new sources an institutional setup that reaches across individual of emissions such that new sources are granted permits jurisdictions – an airshed-based3 management approach. to establish and operate only if they agree to offset every Because air pollution travels across administrative unit of emission by reduction of two units of emission boundaries, and pollution sources are located both elsewhere, a requirement that imposes a heavy cost on inside and outside any given city, an airshed-based new facilities and discourages development. The US management approach that cuts across jurisdictions Clean Air Act also provides a “carrot”: section 105 of the is essential to achieving results. In other words, to Act authorizes the federal government to provide grants effectively address the sources of pollution, air quality equal up to 60 percent of the cost of state air quality should be managed at the same scale as the problem. management programs. Currently, federal funds on The Jing-Jin-Ji (JJJ) Regional Air Quality Prevention and average provide 25 percent of the funding needs of state air Control Coordination Group was established in China programs. Finally, the recently announced performance- to achieve cross-jurisdictional coordination. The group based grants to Indian cities as part of India’s Fifteenth has high-level participation from all administrative Finance Commission recommendations is a step in the entities in the JJJ region, including the Beijing City right direction to create a mechanism to incentivize cities Governor, Tianjin City Governor, and Hebei Provincial to act. Governor, as well as leading officials from the relevant sectoral ministries, including the Ministry of Housing While enforcement of regulations is essential, it is and Urban Development, Ministry of Transportation, not enough; incentives should also be provided to Ministry of Agriculture, and so on. The group is led by the support compliance, and these can entail substantial State Council, China’s highest governmental body. The fiscal outlays. In China, between 2013 and 2017, the 3 An airshed is a part of the atmosphere that behaves in a coherent way with respect to the dispersion of emissions.  6 | CLEARING THE AIR: A TALE OF THREE CITIES group is responsible for formulating targets and annual agricultural emissions, and dust – do not fall under the implementation plans for air quality management across direct purview of pollution boards. Programs such as the JJJ region and setting policies for cross-jurisdictional the Pradhan Mantri Ujjwala Yojana that has expanded issues such as fuel standards, energy supply, and public access to clean cooking fuels, most notably LPG, thereby transportation. Provincial and city-level governments reducing the reliance on residential biomass burning, is continued to be the primary implementers for these air essential to efforts to reduce air pollution. Such a program quality management programs, however. A similar role is goes well beyond the mandate of a pollution control played by the Megalopolis Environmental Commission board and gets to the heart of how development programs in Mexico, which brings together federal authorities from are designed. Similarly, reducing emissions from power the ministries of environment, health, and transport with generation and small and medium enterprises will entail local authorities from Mexico City and 224 municipalities increasing the use of natural gas and renewable energy from the neighboring states of Mexico, Hidalgo, Morelos, and goes well beyond the provisions of the Indian Air Act Puebla, and Tlaxcala, which jointly define an airshed for to prescribe and enforce emission standards for power Mexico City. plants and industry. In China, for example, the ministries of Environmental Protection (now the Ministry of Air pollution management strategies need to be Ecology and Environment), Industry and Information integrated into multi-sector development plans, Technology, Finance, Housing and Rural Development, and an institutional set up is similarly required to along with the National Development and Reform facilitate this, to match the cross-sectoral nature of Commission and National Energy Administration, joined the air pollution challenge. To be effective, air quality together to issue a five-year action plan for air pollution management activities need to be embedded in national prevention and control for the entire JJJ airshed. Their and state development plans, and not just in standalone joint efforts led to the dramatic reduction in coal use in air quality management plans. Notably, three significant the JJJ region. sources of pollution in India – residential biomass burning, CLEARING THE AIR: A TALE OF THREE CITIES | 7 Air Pollution: A Global Challenge Air Pollution: A global challenge 1. Air pollution is one of the leading risks to public to concentrations of PM2.5 pollution higher than 10 health. One of the most dangerous forms of air pollution is micrograms per cubic meter (µg/m3) on average each year, very fine particulates that are capable of penetrating deep or 25 µg/m3 on average every 24 hours (WHO 2005). The into the lungs and entering the bloodstream. Known as available data – drawn from a combination of monitors PM2.5, these particulates have an aerodynamic diameter of measuring PM2.5 on the ground and satellites observing less than 2.5 microns—about one-thirtieth the width of a aerosols from space -- indicate that pollution is far above human hair.4 PM2.5 comes in many forms, including dust, healthy limits in many parts of the world.7 Of the 2,602 dirt, smoke, vapors, gases, microscopic liquid droplets, cities and towns in 89 countries for which the WHO has and heavy metals and comes from a variety of sources. compiled ground-monitored data for average annual Some of the most common sources include emissions PM2.5, about 58 percent (1,517) had concentrations above from burning fossil fuels such as coal or oil and solid the WHO’s guideline, including 97 percent of the 581 biomass such as wood, charcoal, or crop residues. PM2.5 cities and towns in low- and middle-income countries can also come from windblown dust, including natural with data (WHO 2018).8 More broadly, incorporating dust as well as dust from construction sites, roads, and satellite data to measure exposure in areas for which industrial plants.5 Apart from direct emissions, PM2.5 can monitoring does not yet exist, Shaddick et al. (2018) be formed indirectly (known as secondary PM2.5) from estimate that about 94 percent of the world’s people chemical reactions of other pollutants such as ammonia reside in areas for which average annual PM2.5 exceeded (NH3) interacting with sulfur dioxide (SO2) and nitrogen the WHO guideline value in 2015, although the severity oxides (NOx). Exposure to PM2.5 from any or all of these of air pollution varies across these areas.9 sources can cause such deadly illnesses as lung cancer, stroke, and heart disease. Worldwide, an estimated 4.13- 3. Air quality trends have been mixed across the 5.39 million people died prematurely in 2017 from world since the 1990s, with some areas experiencing exposure to PM2.5 pollution—more than the number improvement and others deterioration. The satellite of people who died from HIV/AIDS, tuberculosis, and data also indicate the extent to which air pollution malaria combined.6 has worsened in some regions and improved in others in recent decades. Estimates for average annual PM2.5 2. In many parts of the world, the latest available concentrations in 1990 and 2015 are illustrated in data show that air pollution is far above levels that Figure 1. Table 1 shows average population-weighted are considered healthy. The WHO has recommended exposure by region. Countries in the Middle East and as a guideline that people should not be exposed North Africa, Sub-Saharan Africa, and South Asia have 4 Other pollutants that are commonly monitored to assess the quality of air include PM10 (particulates with diameter 10 microns or smaller), Carbon Monoxide (CO), Nitrogen Dioxide (NO2), Sulphur Dioxide (SO2), Lead (Pb), and Ozone (O3). 5 Although dust may sound benign, the available evidence supports that it is still harmful to people’s health. The current practice of the WHO, the International Agency for Research on Cancer, the U.S. Environmental Protection agency in evaluating the health impacts of PM2.5 is to include dust together with other kinds of particulates (EPA 2009; IARC 2013; WHO 2013). 6 The range represents the 95-percent uncertainty interval. Estimates are from the Global Burden of Disease Study 2017 and include exposure to ambient PM2.5 pollution as well as PM2.5 in households cooking with solid fuels (Stanaway et al. 2018). 7 Beyond the official monitoring in cities, satellite data can help fill the gaps in creating a more complete picture of people’s exposure over time and across the entire country, including in both urban and rural areas. Satellite-flown instruments measure aerosol optical depth (AOD)—the extent to which light reflecting off the Earth’s surface is scattered by particulates and other aerosols in the atmosphere. The data on AOD are then translated into estimates of surface concentrations of PM2.5 using numerical models of atmospheric chemistry and transport. The estimates are calibrated against measurements of PM2.5 (and PM10) at monitoring stations. 8 Data are for the latest year available for measured PM2.5 by city or town. 9 Estimates are based on gridded data for average annual PM2.5 estimated for the Global Burden of Disease Study 2016, as provided to the authors by the Institute for Health Metrics and Evaluation. These satellite-based estimates have been calibrated using ground measurements of PM from more than 6,000 stations in 117 countries. CLEARING THE AIR: A TALE OF THREE CITIES | 9 experienced the most pronounced deterioration in 4. Poor air quality in many regions is taking a heavy air quality over this time period. North America is the toll on the health of populations, as the number only region that meets WHO guidelines for pollution of deaths each year attributed to PM2.5 exposure concentration levels. continues to rise. In 2017, exposure to PM2.5 pollution FIGURE 1: Mean annual ambient PM2.5 pollution across the world, 1990-2015 Mean annual PM2.5 in 1990 Mean annual PM2.5 in 2015 Change from 1990 to 2015 Mean annual PM2.5 (top) (micrograms per cubic meter) 0-10 (WHO guideline) 50 - 75 10 - 35 (WHO IT-1) 75 - 100 35 - 50 Above 100 -15 - 0 (decrease) 10 - 20 0-5 20+ (increase) Sources: IHME (2017), van Donkelaar et al. (2016), 5 -10 Shaddick et al. (2018) TABLE 1: Mean annual population-weighted exposure to ambient PM2.5 per region, 1990-2015 (micrograms per cubic meter, mean estimate and 95-percent uncertainty interval) Region 1990 2005 2015 East Asia and Pacific 38 (25-58) 44 (29-65) 43 (28-63) Europe and Central Asia 24 (13-43) 17 (10-29) 19 (11-33) Latin America and Caribbean 21 (12-34) 22 (13-35) 18 (11-28) Middle East and North Africa 52 (27-95) 50 (25-93) 77 (39-143) North America 11 (7-16) 10 (7-15) 9 (6-13) South Asia 60 (38-91) 66 (42-99) 77 (49-116) Sub-Saharan Africa 49 (21-106) 42 (17-92) 65 (26-142) Notes: Values in parentheses indicate the 95-percent uncertainty interval. Source: Global Burden of Disease Study 2016 data provided by Institute for Health Metrics and Evaluation (IHME); estimates by IHME (2017), van Donkelaar et al. (2016) , and Shaddick et al. (2018). 10 | CLEARING THE AIR: A TALE OF THREE CITIES BOX 1: Air Quality Trends in India Countries usually establish air quality standards for pollutants to define “acceptable” levels of air pollution. India’s National Ambient Air Quality Standards (NAAQS) establish a legal limit of 40 µg/m3 for average annual PM2.5 and 60 µg/m3 for 24-hour concentrations. Monitoring of PM2.5 in India’s cities began only within the past three to four years, making it difficult to assess the full extent and severity of PM2.5 pollution. As of 2017, only around 150 of the country’s some 4,000 cities and towns were covered and monitoring of PM2.5 in rural areas is virtually non-existent, though efforts are underway to expand the network. In 2017, average population-weighted PM2.5 concentrations in the FIGURE B1.1: Monitored PM2.5 pollution in 47 Indian cities, January-November 2018 Ghaziabad, Uttar Pradesh Bhiwadi, Rajasthan Lucknow, Uttar Pradesh Patna, Bihar Faridabad, Haryana Muzaffarpur, Bihar Delhi, Delhi Jodhpur, Rajasthan Moradabad, Uttar Pradesh Kanpur, Uttar Pradesh Agra, Uttar Pradesh Noida, Uttar Pradesh Varanasi, Uttar Pradesh Gaya, Bihar Singrauli, Madhya Pradesh Ahmedabad, Gujarat Rohtak, Haryana Pali, Rajasthan Siliguri, West Bengal Howrah, West Bengal Mandi Gobindgarh, Punjab Asansol, West Bengal Mandideep, Madhya Pradesh Udaipur, Rajasthan Jaipur, Rajasthan Ujjain, Madhya Pradesh Aurangabad, Maharashtra Kota, Rajasthan Amritsar, Punjab Chennai, Tamil Nadu Visakhapatnam, Andhra Pradesh Mumbai, Maharashtra Nagpur, Maharashtra Pithampur, Madhya Pradesh Dewas, Madhya Pradesh Annual mean Ajmer, Rajasthan Ludhiana, Punjab Hyderabad, Telangana 24-hour concentrations Chandrapur, Maharashtra Pune, Maharashtra Indo-Gangetic Plain Rajamahendravaram, Andhra Pradesh Amaravati, Andhra Pradesh Western India Thane, Maharashtra Bengaluru, Karnataka Central India Vijayawada, Andhra Pradesh Thiruvananthapuram, Kerala Southern India Tirupati, Andhra Pradesh Notes: Data for 47 cities are aggregated from 88 automated stations in the national Continuous Ambient Air Quality Monitoring (CAAQM) network; only stations with data for at least 300 days from January to November 2018 are included; the annual mean is averaged from the 24-hour mean concentrations measured at each station in accordance with the national standard. Source: 15-minute and hourly monitoring station data from India’s Central Pollution Control Board, available at http://www.cpcb.gov.in/CAAQM and https://app.cpcbccr.com/ccr/#/caaqm-dashboard-all/caaqm-landing; station data from earlier dates have been collected and archived by OpenAQ, https://openaq.org/. CLEARING THE AIR: A TALE OF THREE CITIES | 11 monitored cities was 73 µg/m3.10 Only one-third of the cities met the national standard for average annual PM2.5 concentrations; none of the locations with continuous monitoring met the national 24-hour standard; and none of the monitored locations met the WHO guidelines. In 2018, the trends remained largely consistent, with the highest levels of PM2.5 being experienced by cities in the Indo-Gangetic Plain (IGP) (Figure B1.1). Average PM2.5 for cities in the IGP during the first 11 months of 2018 ranged from 53 µg/m3 in Ludhiana, Punjab to 161 µg/m3 in Ghaziabad, Uttar Pradesh. Cities in the IGP and western India frequently experienced days when PM2.5 averaged over 200 µg/m3. Pollution levels in cities in southern India tended to be far lower and without the extreme day-to-day variation seen in the IGP , although still higher than acceptable limits. Satellite data provide a more complete picture, across the country and over time, as shown in figure B1.2. These data reinforce the spatial disparities in air quality across the regions but also show that air quality has deteriorated across much of the country since the 1990s. According to these estimates, FIGURE B1.2: Mean annual ambient PM2.5 pollution across the Indo-Gangetic Plain, 1990-2015 Mean annual Mean annual PM2.5 in 1990 PM2.5 in 2015 Mean annual PM2.5 (top) 1990 - 2015 (micrograms per cubic meter) 0-10 (WHO guideline) 75 - 100 10 - 40 (India NAAOS) 100 - 150 40 - 75 150 - 200 Change in mean annual PM2.5 (bottom) (micrograms per cubic meter) -15 - 0 (decrease) 10 - 20 0-5 20+ (increase) Sources: IHME (2017), van Donkelaar et al. (2016), 5 -10 Shaddick et al. (2018) Data are for 72 monitoring stations covering 40 cities that had data for at least 104 days in 2017, the minimum required by India’s National Ambient 10 Air Quality Standard (NAAQS), out of a total of 288 stations and 137 cities in the network as of the end of 2017. Both manual stations in the National Ambient Air Monitoring Programme (NAMP) and automated stations in the Continuous Ambient Air Quality Monitoring (CAAQM) network are included. The population-weighted average is based on the city population as reported in the 2011 census. NAMP and CAAQM monitoring data are provided by the Central Pollution Control Board at http://cpcb.nic.in/manual-monitoring/ and https://app.cpcbccr.com/ccr/#/caaqm-dashboard-all/ caaqm-landing. 12 | CLEARING THE AIR: A TALE OF THREE CITIES the average level of outdoor or “ambient” PM2.5 pollution to which people are exposed each year rose from 39-89 µg/m3 in 1990 to 49-112 µg/m3 in 2015 (IHME 2017; van Donkelaar et al. 2016; Gakidou et al. 2017; Shaddick et al. 2018).11 Anywhere from 732 million to 1.270 billion people in India (56-97 percent of the country ’s population) were exposed to ambient PM2.5 in 2015 above the NAAQS, up from 339-829 million in 1990. Some regions have seen improvements in air quality (areas in green in the bottom panel of Figure B1.2), though concentrations remain high and above healthy levels even in these areas. Finally, poor air quality is not only an urban issue. As Figure B1.2 shows, air pollution particularly in the IGP and a few other regions is widespread and affects rural communities as well as people in cities. resulted in 4.13-5.03 million premature deaths globally, 5. The health impacts of air pollution also represent including about 2.50-3.36 million deaths from outdoor a heavy cost to the economy. Public agencies in many ambient PM2.5 and 1.40-1.93 million deaths from PM2.5 countries have taken a variety of approaches to quantifying in households cooking with solid fuels (Balikrishnan et the economic cost of air pollution and the benefits of al. 2018; Stanaway et al. 2018).12 In other words, about policies aimed at reducing pollution. One of the most 7.4-9.0 percent of attributable deaths globally in 2017 widely accepted approaches applied by governments were linked to PM2.5 pollution (Figure 2), more than were is based on individuals’ expressed willingness to pay to caused by HIV/AIDS, tuberculosis, and malaria combined. reduce their risk of dying.13 Following this approach, the Furthermore, about half of the deaths attributed to PM2.5 economic cost of fatal illness caused by PM2.5 pollution pollution occurred among people younger than 70 years. globally in 2017 was on the order of US$ 2.248 trillion FIGURE 2: Leading fatal health risks globally, 2017 Metabolic risks 31.4 Dietary risks 19.5 Tobacco 14.5 Air pollution (PM2.5) 8.2 Child and maternal malnutrition 5.7 WASH 2.9 Low physical activity 2.3 0 5 10 15 20 25 30 35 Share of all-cause mortality globally in 2017(%) Notes: “Air pollution (PM2.5)” includes ambient PM2.5 pollution and household PM2.5 pollution from cooking with solid fuels; “WASH” includes unsafe water, sanitation, and handwashing. Source: Institute for Health Metrics and Evaluation, Global Burden of Disease Study 2017 (2018); Stanaway et al. (2018). 11 The ranges presented here represent 95-percent uncertainty intervals, with a central estimate of 60 µg/m3 for 1990 and 76 µg/m3 for 2015. These estimates of ambient PM2.5 exposure are from the Global Burden of Disease Study 2016 (GBD 2016), an international scientific effort led by the Institute for Health Metrics and Evaluation at the University of Washington, Seattle, United States 12 The ranges represent 95-percent uncertainty intervals. Deaths from exposure to PM2.5 pollution as estimated for the Global Burden of Disease Study 2017 (GBD 2017) and reported here included deaths from acute lower respiratory infections, diabetes, ischemic heart disease, stroke, chronic obstructive pulmonary disease, and lung cancer (Balikrishnan et al. 2018; Stanaway et al. 2018). CLEARING THE AIR: A TALE OF THREE CITIES | 13 to US$ 13.692 trillion (95-percent uncertainty interval), loss of income loss due to fatal illness from PM2.5 pollution with a central estimate of US$ 4.334 trillion.14 The wide in 2017 would be in the range of US$ 131-317 billion range accounts for uncertainty owing to the health globally, with a central estimate of US$ 200 billion, equal impacts as well as people’s willingness to pay. In other in magnitude to 0.1-0.3 percent of GDP.15 words, even under the most conservative scenario, the annual economic cost of PM2.5 pollution is equivalent 6. Apart from the cost of fatal illness, air pollution in magnitude to at least 1.9 percent of the world GDP. impacts a country’s economy in other ways too. (Costs are expressed as an equivalent percent of GDP Ground-level ozone (O3) pollution, for example, which just to provide a convenient sense of scale, not to forms when volatile organic compounds (VOCs) react suggest they are a direct loss of GDP. GDP is a measure with NOx, is toxic to plants and has been shown to reduce of output, not economic wellbeing.) As an alternative, crop yields. One study found that ozone and black other governments have also measured the loss of human carbon (emitted mostly from household cookstoves) capital due to fatal illness. Under this approach, losses are cut yields of wheat and rice by about 33 percent and 22 estimated in terms of the expected labor income that percent, respectively, in India’s largest producing states people would have earned over their lifetimes had they in 2010. Lower yields translated into a loss of 24 million not died prematurely. The estimated loss of income is tons of harvested wheat alone, worth about US$ 5 billion typically much smaller than the total economic cost of (Burney and Ramanathan, 2015). Elsewhere, research fatal illness as estimated based on individuals’ willingness from China shows that air pollution is also making to pay, reflecting how people value more than just their skilled workers in urban areas less productive (Chang et paychecks. Alternatively, if losses are calculated only on al. 2016), suggesting that worsening air quality may be the basis of forgone lifetime labor earnings, the expected dulling the competitive edge of cities too. 13 Economists use a variety of methods to elicit people’s willingness to pay (WTP). One method is through so-called stated preference surveys. When people who are surveyed tell economists how much they are willing to pay to reduce their fatality risk, they are thinking about much more than their paychecks. Losses may also reflect the loss of enjoyment that people get from intangibles such as being alive or spending time with loved ones. Another method is by looking at the differences in wages for more or less risky jobs. The estimates of the costs of pollution presented here are based primarily on findings from stated preference surveys. For a discussion of why, please refer to Narain and Sall (2016). 14 Monetary losses are reported in terms of US dollars calculated at constant year 2011 prices on a purchasing power parity (PPP) basis. Global welfare losses here represent the sum of losses as calculated for 168 countries. See World Bank-IHME (2016) for a description of how the uncertainty interval is calculated based on varying key assumptions such as WTP for reduced fatality risk using a database of WTP estimates from stated preference studies around the world. The central estimate represents the median estimate from 5,000 random draws. 15 Global estimates presented here are for 164 countries. Fewer countries have the necessary data to calculate forgone labor output than welfare losses. As above, see World Bank-IHME (2016) for a description of how the uncertainty interval is constructed. 14 | CLEARING THE AIR: A TALE OF THREE CITIES What is the relationship between air quality and economic growth? What is the relationship between air quality and economic growth? 1. Low income and lower-middle income countries (LICs and LMCs) have experienced deteriorating air FIGURE 3: Mean annual ambient PM2.5 pollution versus GNI quality since 1990, though not upper-middle and per capita, 1990 and 2015 high-income countries (see Table 2). Between 1990 and 2015, mean annual PM2.5 in LICs increased from 160 44 µg/m3 to 56 µg/m3, while mean annual PM2.5 in the Mean annual PM2.5 LMCs (excluding India) rose from 45 µg/m3 to 59 µg/ 80 1990 m3. In India, during this time period, mean annual 40 2015 PM2.5 also increased, from an average of 60 µg/m3 in 20 1990 trend 1990 to 76 µg/m3 in 2015. At the same time, ambient 2015 trend 10 PM2.5 stabilized in some upper-middle-income – China 5 -- and improved in others – Mexico, for example -- and 250 1,000 4,000 16,000 64,000 in high-income countries. These trends imply that GNI per capita (constant 2010 US$) the disparity in air quality between poorer and richer countries has grown over time, as further illustrated in Notes: Mean annual PM2.5 is population-weighted mean exposure; PM2.5 and GNI per capita are rescaled logarithmically. Figure 3. In the figure, blue dots represent mean annual Source: Global Burden of Disease Study 2016 data provided by Institute PM2.5 concentrations at different levels of per capita for Health Metrics and Evaluation (IHME); estimates by IHME (2017), income in 2015, and orange dots the same in 1990. A van Donkelaar et al. (2016), and Shaddick et al. (2018); GNI per capita data from World Bank, World Development Indicators database. steeper trend line in 2015 illustrates the growing disparity. TABLE 2: Mean annual population-weighted exposure to ambient PM2.5 per income group and for China, India, and Mexico, 1990- 2015 (micrograms per cubic meter, mean estimate and 95-percent uncertainty interval) Income Group or Country 1990 2005 2015 Low income 44 (15-106) 40 (14-96) 56 (19-137) India (lower middle income) 60 (39-89) 66 (42-98) 76 (49-112) Other lower middle income 45 (24-79) 44 (24-76) 59 (31-104) China (upper middle income) 48 (31-72) 57 (38-83) 56 (38-82) Mexico (upper middle income) 25 (15-37) 26 (16-39) 19 (12-28) Other upper middle income 26 (15-45) 23 (13-39) 25 (14-44) High income 18 (11-26) 16 (10-23) 19 (12-29) Notes: Values in parentheses indicate the 95-percent uncertainty interval; data for China, India, and Mexico are highlighted in grey; income group classifications are as of the 2015 calendar year. Source: Global Burden of Disease Study 2016 data provided by Institute for Health Metrics and Evaluation (IHME); estimates by IHME (2017), van Donkelaar et al. (2016), and Shaddick et al. (2018). 16 | CLEARING THE AIR: A TALE OF THREE CITIES 2. These trends would seem to reinforce that the FIGURE 4: GDP per capita versus mean annual experience of any country is part of a larger trend ambient PM2.5 in select large middle-income countries, consistent with the so-called “Environmental Kuznets 1990-2015 Curve” . Named after Simon Kuznets, who theorized that income inequality initially worsens and then improves at 80 Nepal Pakistan India higher levels of income as a country develops (Kuznets, 70 1955), the notion that pollution and income also form 60 an inverted U-shaped curve was first popularized in Mean annual PM2.5,1990-2015 (μg/m3) China the early 1990s and is referred to as the Environmental 50 Myanmar Iran, Islamic Rep. Kuznets Curve (EKC).16 The possibility that a country’s 40 development path follows an EKC, raises the question South Africa Turkey 30 Sri Lanka whether poor air quality is a reflection of a country’s low Vietnam level of development. Will countries simply grow their 20 Ukraine Thailand Mexico Kazakhstan way out of pollution? 10 Brazil 3. Large middle-income countries appear to be on 0 0 5,000 10,000 15,000 20,000 25,000 very different growth paths when assessed by the GDP per capita, 1990-2015 (year 2011 US$, PPP) level of pollution intensity. Figure 4 illustrates the relationship between mean annual PM2.5 exposure and Source: Global Burden of Disease Study 2016 data provided by Institute for Health Metrics and Evaluation (IHME); estimates by the level of income, as measured by GDP per capita, for IHME (2017), van Donkelaar et al. (2016), and Shaddick et al. (2018); large middle-income countries from 1990 to 2015. A few GDP per capita data from World Bank, World Development Indicators database. countries – China and Mexico among them – appear to follow the inverted U-shaped curve, where pollution first 4. Trends at the state-level within India also suggest increases as the country grows. These countries appear that economic growth has been much more pollution- to have reached a turning point, though at different intensive in some parts of the country than in others. levels of income per capita, after which pollution started Figure 5 compares GDP per capita in purchasing power to fall. Number of other countries are yet to reach a parity (PPP) terms with mean annual exposure to PM2.5 in turning point and mean annual PM2.5 is increasing in the various states and regions of India. Trends for China these countries along with income. Among this group and Mexico are shown for comparison. As the figure of countries, countries in South Asia – India, Nepal, reveals, India’s pollution-intensive growth pattern is and Pakistan – stand out and appear to have a more driven primarily by trends in states of the IGP , including pollution-intensive growth path than countries of similar Bihar, Delhi, Haryana, Jharkhand, Punjab, Uttar Pradesh, income levels, where pollution intensity is measured and West Bengal. These seven states had the highest by the steepness of the curve. Econometric analysis elasticities of PM2.5 with respect to income, although at further reveals that income elasticity of ambient PM2.5 is markedly different levels of per capita income: Bihar and systematically higher in the South Asia region than for Jharkhand had the lowest levels of GDP per capita in other regions, even after accounting for differences in the country, while Delhi and Haryana had the highest. economic structure, demographics, energy supply, and Other than states in the IGP , states in Central India, and natural characteristics such as topography and climate Western states appear to be on more pollution-intensive (see Annex). development paths. Box 2 explores the reasons underlying these somewhat paradoxical trends. One of the first empirical studies to suggest the existence of an EKC was a 1991 paper by Grossman and Krueger. Responding to concerns that 16 environmental quality in Mexico might suffer if its economy was opened to polluting industries under the North American Free Trade Agreement, Grossman and Krueger (1991) tested how air quality in a cross-section of cities in more than 40 countries varied across different levels of GDP per capita and trade openness (exports as a share of GDP). They found that average concentrations of SO2 and smoke tended to increase with GDP per capita in the lowest-income countries but then fell as GDP continued to increase beyond a certain level. CLEARING THE AIR: A TALE OF THREE CITIES | 17 FIGURE 5: Mean annual PM2.5 pollution and GDP per capita in various regions of India, 1990-2015 a. Indo-Gangetic Plain (plus China and Mexico) b. Central States 160 160 Delhi 140 140 120 Bihar 120 Mean annual PM2.5 (μg/m3) Mean annual PM2.5 (μg/m3) Uttar Pradesh Haryana 100 West Bengal 100 Punjab Jharkhand 80 80 Madhya Pradesh 60 60 Odisha China China Maharashtra 40 40 Chhattisgarh 20 20 Mexico Mexico 0 0 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 GDP per capita (2011 US$, PPP) GDP per capita (2011 US$, PPP) c. Southern States d. Northeastern States 160 160 140 140 120 120 PM2.5 (μg/ PM2.5 (μg/ 100 100 80 80 m3) Mean annual m 3) Mean annual 60 60 Tripura China Assam China Telangana Meghalaya Andhra Pradesh 40 40 Tamil Nadu Sikkim Kerala Manipur Arunachal Pradesh 20 Karnataka Mexico 20 Mizoram Mexico 0 0 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 GDP per capita (2011 US$, PPP) GDP per capita (2011 US$, PPP) e. Western Himalayan States f. Western States 160 150 140 125 120 Mean annual PM2.5 (μg/m3) Mean annual PM2.5 (μg/m3) 100 100 80 75 Rajasthan Gujarat 60 Uttarakhand China China Jammu and Kashmir 50 Himachal Pradesh Assam 40 25 20 Mexico Mexico 0 0 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 GDP per capita (2011 US$, PPP) GDP per capita (2011 US$, PPP) Notes: “PPP” = purchasing power parity. Sources: Ambient PM2.5 exposure data from Global Burden of Disease Study 2016, as provided by IHME; estimates by IHME (2017), van Donkelaar et al. (2016) , and Shaddick et al. (2018); gridded GDP data from Kummu et al. (2018). 18 | CLEARING THE AIR: A TALE OF THREE CITIES BOX 2: Explaining two paradoxes about pollution intensity of India’s growth path State-level trends of air pollution and economic growth shown in Figure 5 reveal two seeming paradoxes about India’s growth path. First, how can some states with the lowest GDP per capita and slowest rates of economic growth also suffer from the worst air pollution? Such high levels of air pollution are commonly associated with the processes of industrialization and urbanization, which tend to occur at relatively higher levels of income per capita than the current levels in the poorest states of the IGP. Second, how can states such as Delhi, Haryana, and Punjab have high levels of pollution despite being relatively rich? As shown in Figure 4, countries at income levels comparable to these three states were able to improve their air quality while continuing to grow. Why have these three states not been able to do the same? Recent studies examining how emissions from different sectors have contributed to overall ambient concentrations of pollution in the air that people breathe across Indian states reinforce the cross-sector and cross-border nature of pollution in India, offering an explanation for the paradoxes. Analysis by the Council on Energy, Environment and Water (CEEW) and International Institute for Applied Systems Analysis (IIASA) sheds light on the contribution of different sectors and different regions to poor air quality in a particular region. Figure B2.1 shows that residential emissions from households using solid fuels for heating and cooking, emission from industry, power plants, agriculture, and transport, and natural dust all contribute to poor air quality, though to differing levels in different states. It also points to the significant contribution of secondary PM2.5 formed by reactions involving other pollutants. Because PM2.5 can remain suspended in the atmosphere for FIGURE B2.1: State-wise origin of mean annual population-weighted ambient PM2.5 by sector, 2015 120 100 80 PM2.5 (μg/m3) 60 NAAQS 40 WHO 20 guideline 0 Kerala Goa Delhi West Bengal Haryana Uttar Pradesh Jharkhand Bihar Punjab Gujarat Odisha Rajasthan Chhattisgarh Maharashtra Madhya Pradesh Uttarakhand Andhra Pradesh* North East Assam Karnataka Tamil Nadu Jammu, Kashmir Himachal Pradesh Natural sources Secondary PM2.5* Coal thermal power plants Other high stacks Households Transport Waste Agriculture Source: CEEW-IIASA (2018). CLEARING THE AIR: A TALE OF THREE CITIES | 19 FIGURE B2.2: Contribution of local, regional, and transboundary sources of ambient PM2.5 exposure in Indian states, 2015 (micrograms per cubic meter) 120 100 80 PM2.5 (μg/m3) 60 40 NAAQS WHO 20 guideline 0 Madhya Pradesh Rajasthan Chhattisgarh Karnataka Haryana Gujarat Odisha Andhra Pradesh* Punjab Uttarakhand Goa Maharashtra Bihar Jharkhand North East Himachal Pradesh Uttar Pradesh Delhi West Bengal Assam Kerala Tamil Nadu Jammu and Kashmir Natural sources Outside India Other India Neighboring States This States Source: CEEW-IIASA (2018). long periods of time and travel hundreds or thousands of kilometers, PM2.5 emissions can cross state boundaries. Figure B2.2 captures this characteristic feature of PM2.5 concentrations finding that states and Union Territories (UTs) with the greatest amount of ambient PM2.5 originating from other states and regions in India include Bihar, Delhi, Haryana, Jharkhand, and Odisha. These results suggest that individual states acting in isolation are unlikely to solve their air quality problems on their own. The high levels of pollution at relatively low levels of development in four IGP states and the high level of pollution in the three IGP states with the highest GDP per capita can be explained by a combination of factors that follow from the contribution of the different sources to air pollution and regional nature of the challenge. Agricultural intensification, high population density, a high reliance on residential biomass, and regional sources of PM2.5 from other states and areas help explain the high levels of pollution in the IGP at relatively low levels of economic development. Low levels of development are characterized by a high share of agriculture in the economy and certain consumption patterns, such as a high degree of dependence on biomass for cooking and heating. In fact, most of India’s IGP is covered by cropland (about 85 percent of the land area), scattered with a few large (7) and medium-sized (22) cities and many smaller cities (37) and towns (648). The agriculture sector is an important contributor to the economy as well as PM2.5 pollution, with intensive farming, high fertilizer use, and the seasonal burning of crop residues. Consequently, Chakraborty and Gupta (2010) found 20 | CLEARING THE AIR: A TALE OF THREE CITIES that secondary sources (with NH3 from agriculture reacting with NOx and SO2) contributed about 40 percent of PM2.5 concentrations in the city of Kanpur in Uttar Pradesh. High population density in rural areas compounds the problem, with households continuing to depend primarily on biomass for cooking and heating. Apart from Tamil Nadu and Kerala, the states in the IGP have the highest population density of any in the country. High population density and low levels of development also make small township-based industries an important source. Finally, due to features of the IGP’s geography and climate, the dispersion of pollution is weak, particularly in the winter months, and pollution from upwind areas is funneled into the region, adding to the challenge. The same factors, combined with greater emissions from motor vehicles, further help to explain why India’s states with the highest levels of GDP per capita—Delhi, Haryana, and Punjab—also experience some of the worst pollution. Studies in the Delhi National Capital Region (NCR) have found that vehicular emissions contribute about 23-25 percent of ambient PM2.5 in the winter and about 19 percent of mean annual ambient PM2.5. Also, although cooking and heating with liquified petroleum gas (LPG), electricity, and natural gas is more common in urban districts of the NCR and surrounding areas, the three states also continue to have a high density of households reliant on solid fuels. Furthermore, as in the other states in the IGP , agriculture continues to be an important economic sector for Punjab and Haryana, with farming characterized by greater intensification, higher fertilizer use, and the seasonal burning of crop residues. Regional sources of air pollution also contribute a large share of overall ambient PM2.5 concentrations, with as much as 60 percent of Delhi’s pollution coming from neighboring states (Amman et al. 2016). 5. The high pollution intensity of economic growth growth higher than 3 percent between 1990 and 2015 saw for South Asian countries is not the norm. Many low- air quality deteriorate as much as it did in India. All the and middle-income countries that have experienced countries with GDP per capita growth rates higher than rapid income growth have achieved a reduction in India saw smaller increases or decreases in ambient PM2.5 ambient PM2.5. Only six of the 42 low and middle-income (see Figure 6). countries with average annual rates of GDP per capita FIGURE 6: Mean annual PM2.5 in 2015 (left) and change in mean annual PM2.5 from 1990 to 2015 (right) for low- and middle- income countries with average annual GDP per capita growth of at least 3 percent 140 40 Bangladesh Change in mean annual PM2.5 in 1990-2015 (μg/m3) Nigeria 120 30 Nigeria Chad Mean annual PM2.5 in 2015 (μg/m3) Iraq 100 Bangladesh 20 Sudan India Chad Uganda Ethopia 10 Mongolia 80 Armenia Sudan India Uganda Iraq Malasia Georgia Lao PDR Azerbaijan China Cambodia Myanmar 60 0 Vietnam Rwanda China Thailand Albania Ethopia Cambodia Uzbekistan Sri Myanmar Serbia Bosnia and Uzbekistan VietnamBosnia and Lanka Herzegovina 40 Turkmenistan Angola -10 Herzegovina Romania Belarus Mongolia Lao PDR Azerbaijan Moldova Armenia 20 Mozambique -20 Bulgaria Sri Albania Lanka 0 -30 2 3 4 5 6 7 8 9 10 2 3 4 5 6 7 8 9 10 Average GDP per capita growth rate, 1990-2015 (%) Average GDP per capita growth rate, 1990-2015 (%) Source: Global Burden of Disease Study 2016 data provided by Institute for Health Metrics and Evaluation (IHME); estimates by IHME (2017), van Donkelaar et al. (2016), and Shaddick et al. (2018); GDP per capita data from World Bank, World Development Indicators database. CLEARING THE AIR: A TALE OF THREE CITIES | 21 6. These trends reinforce the role of policies and foundation. Policy decisions, investments, technologies, programs in affecting the shape of the relationship and external shocks can bend, flatten, or shift the curve between economic growth and pollution, suggesting (Payanotou 1997; Unruh and Moomaw 1998; Dasgupta that countries cannot simply grow their way out of et al. 2002), meaning that countries cannot simply rely pollution. Although the existence of the EKC continues on economic growth to produce a cleaner environment to be debated in the academic literature (see Stern 2015 for in the end.17 a recent review), critics have rightly questioned its shaky Indeed, this point was raised by one of the earliest EKC studies, published in 1992 by World Bank economists Nemat Shafik and Sushenjit Bandyopadhyay, 17 and is worth revisiting. Shafik and Bandyopadhyay tracked the relationship of income, investment, trade, debt and other macroeconomic characteristics with a variety of indicators for environmental quality, including mean annual concentrations of TSP in cities. Looking at trends in urban air quality across countries from the 1970s to the 1980s, they found evidence of an inverted U-shaped relationship of TSP with income—like that observed for SO2 and smoke by Grossman and Krueger (1991). They conclude: “The evidence suggests that it is possible to ‘grow out of’ some environmental problems. But there is nothing automatic about this—policies and investments must be made to reduce degradation” (Shafik and Bandyopadhyay 1992: 23). As they note, not all indicators of environmental quality improved at higher income levels: carbon emissions and municipal waste per capita, for example, increased exponentially. 22 | CLEARING THE AIR: A TALE OF THREE CITIES Tackling air pollution: Lessons from Mexico City, Beijing, and Delhi Tackling air pollution: Lessons from Mexico City, Beijing, and Delhi 1. Role of policies, technologies, and investments from the Governments of the Federal District and the Come to light when one looks at the experience of State of Mexico, municipal governments in the greater many countries, and urban agglomerations within metropolitan area, the Federal Electricity Commission, them, that have successfully tackled the air pollution the Mexican Institute of Petroleum, and the Mexican challenge. Three cities -- Mexico City, Beijing, and Delhi state-owned petroleum company -- Pemex. The working – offer lessons for other cities as they take on the fight to group developed the Integrated Program against reduce air pollution. Atmospheric Pollution in the Mexico City Metropolitan Area (PICCA) for 1990-1994 (GoM 1990), the first air quality management strategy. 3A. AIR QUALITY MANAGEMENT IN MEXICO CITY 4. A number of measures were implemented as part of the PICCA to reduce air pollution. Under 2. Mexico City suffered from hazardous air pollution the PICCA, pollution abatement measures were in the 1980s and 1990s, as a result of rapid population introduced in the greater metropolitan area, which then growth, industrialization, and motorization. In encompassed Mexico City and 17 municipalities from 1992, a comparison of key pollutants – lead (Pb), sulfur the State of Mexico. The biggest achievements were: dioxide (SO2) and total suspended particles (TSP) -- improving fuel quality (phasing out of lead in gasoline, across several mega cities in Mexico led the World Health introduction of oxygenated gasoline, introduction of Organization and the United Nations Environment diesel with a maximum sulfur content of 500 PPM, and Program to conclude that Mexico City was the most substitution of heavy fuel oil with light fuel oil in 1991 polluted megalopolis on the planet (WHO-UNEP 1992). and gasoil in 1995), installation of pollution control Monitoring data for these pollutants and nitrogen oxide technologies in vehicles (three way catalytic converters, (NOx) and ozone (O3) collected in the city beginning evaporative emission controls, and electronic fuel in 1986 showed that pollution frequently exceeded injection and ignition accessories), and the establishment national standards and the WHO air quality guidelines of stricter vehicular pollution norms (GoM, n.d.). As (see Figure 7). a result, Pb, carbon monoxide (CO), PM10, and SO2 emissions fell significantly, and O3 emission levels 3. Facing a public health crisis, the Government stabilized (GoM 1996). Pollution declined dramatically of Mexico launched the first multi-year air quality in just a few years (see Figure 7). management strategy in 1990. At this time Mexico City was known as the Federal District and was administered 5. Building on lessons from the implementation of by the Chief of the Federal District, as part of the Federal PICCA, additional programs were introduced to Administration and reporting directly to the President. control air pollution, starting in the mid-90s. PICCA In response to the President’s call to act immediately was followed by the Program to Improve Air Quality to abate air pollution, an inter-agency working group (PROAIRE) 1995-2000, that included for the first time in was established that included representatives from the Mexico quantitative air quality improvement goals. The Ministries of Urban Development and Environment, PROAIRE’s main goal was to reduce peak and average Finance, Planning and Budget, Commerce and Industrial concentrations of ozone, which had not been reduced Promotion, Health, Energy, Mines and State-Owned significantly under PICCA. Though PROAIRE succeeded Industries, Agriculture and Hydraulic Resources, and in reducing O3 levels significantly, concentrations of Transport and Communications, as well as representatives O3 remained above the legal standard for more than 80 24 | CLEARING THE AIR: A TALE OF THREE CITIES FIGURE 7: Monitored concentrations of pollutants in Mexico City compared to national standards and WHO guidelines, January 1986 to September 2018 a. Hourly NO2 b. Mean Annual NO2 100 700 50th percentile 95th percentile Maximum 650 90 600 80 550 500 70 Mean annual NO2 (μg/m3) Hourly NO2 (μg/m3) 450 60 400 National standard 350 50 300 WHO guideline 40 250 WHO guideline 200 30 150 20 100 10 50 0 0 1986 1990 1994 1998 2002 2006 2010 2014 2018 1986 1990 1994 1998 2002 2006 2010 2014 2018 c. Hourly O3 d. 8-hour O3 900 50th percentile 95th percentile Maximum 600 50th percentile 95th percentile Maximum 550 800 500 Ozone, 8-hour running average (μg/m3) 700 450 Hourly ozone (μg/m3) 600 400 500 350 300 400 250 National standard 300 200 National standard 150 200 WHO guideline 100 100 50 0 0 1986 1990 1994 1998 2002 2006 2010 2014 2018 1986 1990 1994 1998 2002 2006 2010 2014 2018 e. Daily (24-hour) PM10 f. Mean annual PM10 450 50th percentile 95th percentile Maximum 80 400 70 PM10, 24-hour mean (μg/m3) 350 60 Mean annual PM10 (μg/m3) 300 50 250 National standard 40 200 30 150 WHO guideline 20 100 National standard 50 10 WHO guideline 0 0 1995 2000 2005 2010 2015 1995 2000 2005 2010 2015 CLEARING THE AIR: A TALE OF THREE CITIES | 25 g. Daily (24-hour) PM2.5 h. Mean annual PM2.5 175 50th percentile 95th percentile Maximum 40 150 PM2.5, 24-hour mean (μg/m3) Mean annual PM2.5 (μg/m3) 30 125 100 20 75 National standard 50 National standard 10 WHO guideline 25 WHO guideline 0 0 2003 2006 2009 2012 2015 2018 2003 2006 2009 2012 2015 2018 i. 3-month Pb j. 24-hour SO2 2.5 50th percentile Maximum 500 50th percentile 95th percentile Maximum , 3-month running average (μg/m3) 450 SO2, 24-hour running average ( (μg/m3) 2 400 350 National standard 1.5 300 National standard 250 1 200 150 100 Lead in TSP .5 50 WHO guideline 0 0 1986 1990 1994 1998 2002 2006 2010 2014 2018 1989 1992 1995 1998 2001 2004 2007 2010 2013 Notes: Data are from Mexico City’s continuous air quality monitoring network (RAMA). To provide a representative picture of trends over time, only stations with data for the full span of months and years are included, leaving three stations for NO2 (MER, PED, XAL); five stations for SO2 (FAC, MER, PED, TLA, XAL); one station for O3 (PED); 10 stations for PM10 (CES, LVI, MER, NET, PED, TAH, TLA, TLI, VIF, XAL); five stations for PM2.5 (CAM, MER, SAG, TLA, UIZ); and four stations for Pb (MER, PED, TLA, XAL). NO2 data are converted from parts per billion (ppb) to micrograms per cubic meter (µg/m3) assuming reference conditions as defined in the national standard, NOM-023-SSA1-1993. SO2 data are for a lagged 24-hour running average, converted from ppb to µg/m3 assuming reference conditions as defined in the national standard, NOM-022-SSA1-1993. Eight-hour O3 data are for a lagged 8-hour running average. O3 data are converted from ppb to µg/m3 assuming reference conditions as defined in the national standard, NOM-020-SSA1-2014. Twenty-four-hour PM10 and PM2.5 data are for the daily 24-hour period, beginning and ending at midnight, as defined in the national standard, NOM-025-SSA1-2014. Pb data are for lagged 3-month running average, as per the national standard, NOM-026-SSA1-1993. Station codes defined in the RAMA and REDMA metadata. Source: Mexico City SEDEMA, https://www.sedema.cdmx.gob.mx/; WHO (2005). 26 | CLEARING THE AIR: A TALE OF THREE CITIES percent of the days in 2000. The next phase of PROAIRE Commission (la Comisión Ambiental de la Megalópolis, (2002-2010) incorporated new scientific information and “CAMe”) in 2013, to cover a greater geographic area, adopted goals to reduce the number of days in which recognizing that the metropolitan area had grown and the concentrations of O3, PM10, SO2, and CO exceeded that the sources of and responses to pollution required respective air quality standards. A national standard for a geographical coverage that was wider than CAM’s. PM2.5 concentrations was also introduced for the first CAMe’s geographical coverage included Mexico City and time in 2004. The interventions implemented under the a total of 224 municipalities in the surrounding states second PROAIRE helped to reduce the emissions of all of Hidalgo, Mexico, Morelos, Puebla, and Tlaxcala. In measured pollutants but annual concentrations of PM10 addition, CAMe had a more comprehensive governance and PM2.5 and daily concentrations for O3 continued to structure that included a high level government body exceed the legal standards. PROAIRE 2011-2020 built (with Mexico’s Minister of Environment and Natural on previous experience, incorporated the latest scientific Resources, the five state governors, and the Chief of Mexico findings, and proposed managing the Mexico City City’s Government), an Executive Commissioner, and a Metropolitan Area (MCMA) as an ecosystem (MESM- Scientific Advisory Committee.19 States that were part MEFD-MENR-MH 2011). Finally, in 2017, Secretariat of CAMe agreed in 2014 to transfer to the commission of Environment and Natural Resources (Secretaria del MX$ 5.00 for each vehicle inspection conducted within Medio Ambiente y Recursis Naturales, “SEMARNAT”) their jurisdiction to complement the trust fund created published the “Federal Program to Improve Air Quality in 1992 to finance new environmental projects. in the Megalopolis: PROAIRE of the Megalopolis 2017- 2030.” (MENR 2017). TABLE 3: Primary Emissions from Mobile Sources in the Mexico City Metropolitan Area (% of total emissions, 6. Coordination between the federal and local selected years) governments has been key to air quality management in the Mexico City Metropolitan Area. In 1992, the Year PM10 PM2.5 SO2 NOx VOC Metropolitan Environmental Commission (CAM) was created by Presidential Decree to improve coordination 1998 36% N/A 21% 81% N/A between federal and local government agencies in the 2004 23% 57% 45% 82% 35% airshed.18 The CAM brought together federal authorities 2008 16% 52% 49% 82% 31% from the ministries of environment, health, and 2016 53% 56% 28% 86% 17% transport and local authorities from the Federal District (now Mexico City) and the State of Mexico. The Federal Source: Emissions Inventory of the Mexico City Metropolitan Area 1998, 2004, 2008, and 2016. Government also established a trust fund that would serve as CAM’s financial mechanism and that was funded by a gasoline charge that was in place between 1995 and 7. Reducing pollution from vehicles has remained a 1997 (NIGSI 2005). In its early years, the CAM struggled top priority of the Mexico City Metropolitan Area to fulfill its role as the coordinating body for air quality (MCMA) pollution management strategy. Emissions management due to frequent personnel changes, lack of inventories published since 1998 showed that mobile an independent budget, lack of enforcement powers, and sources are a key source of primary particle emissions an unclear organizational structure (Molina and Molina in the MCMA and precursors to secondary pollutants, 2006). However, CAM’s gradual strengthening allowed particularly PM2.5 and oxidants like ozone (see Table it to play an instrumental role in the development and 3). Policies aimed at reducing harmful emissions from implementation of PICCA and air quality programs vehicles in the MCMA have combined gradually tighter (PROAIRE) for 1995-2000, 2002-2010, and 2011-2020. standards, improved inspection and maintenance, and The CAM evolved into the Megalopolis Environmental monetary incentives. Air quality management plans for 18 The commission was established by a Presidential Decree published on January 8, 1992 and its original name was the Commission for the Prevention and Control of Atmospheric Pollution in the Metropolitan Area of the Valley of Mexico. The Governments of the Federal District Department, the State of Mexico and the State of Hidalgo signed an agreement on September 13, 1996 to increase the commission’s geographic coverage to the entire Federal District, 59 municipalities in the State of Mexico and 29 municipalities in the State of Hidalgo. The agreement also changed the commission’s name to Metropolitan Environmental Commission (Comisión Ambiental Metropolitana-CAM). 19 Agreement published in the Diario Oficial de la Federación, October 3, 2013 CLEARING THE AIR: A TALE OF THREE CITIES | 27 2002- 2010 and 2011-2020 (“PROAIRE 2002-2010” and Euro VI/US EPA 2010 standards is not scheduled “PROAIRE 2011-2020”) include vehicular emission until 2021. reduction policies such as: < Subsidies and low-cost loans for the replacement < Progressively tighter fuel quality standards: State and upgrading of taxis and buses: By the early- and local authorities have coordinated with the Federal 2000s, there were about 110,000 taxis on the streets Government and PEMEX through the CAM (and later of the greater metropolitan area and most of these CAMe) to advocate for more stringent fuel standards were at least nine years old. Due to their age and the An official norm published in January 2006 (NOM- long distances traveled each day, taxis contributed an 086-SEMARNAT-SENER-SCFI-2005) restricted the oversized share of NOx, CO, VOC, and PM emissions. sulfur content in premium gasoline to a maximum To confront this problem, during the PROAIRE plan of 80 parts per million (PPM) by October 2006 and (2002-2010), the city set a goal of scrapping about 80 required a similar sulfur content for magna gasoline percent of old taxis by 2007. In the Federal District, sold in the MCMA by October 2008. The same norm the city provided direct subsidies of about US$ 1,500 required that only diesel with a maximum sulfur to drivers in exchange for retiring and scrapping content of 15 ppm be sold in the MCMA by January their old vehicles, along with access to low-cost loans 2009.20 In the rest of the country, the use of ultra-low for vehicle refurbishment or the purchase of more sulfur diesel was required starting in December 2018, efficient, less polluting vehicles. Economists estimate bringing Mexico’s fuel quality in line with the highest that by providing subsidies and low-cost loans, public European and US standards (NIECC 2019). authorities were able to leverage private investment by a ratio of about 3:1 and that the benefit of annual < Vehicle emissions standards: Vehicle emissions fuel savings outweighed the public cost by about 6:1 standards have been progressively tightened in (McKinley et al. 2005). step with better fuel quality. Mexico City has anticipated tightening emissions standards and < Improved public transit: During PROAIRE 2002- forged ahead of the national timetable by adopting 2010, the city began operating a new bus rapid transit standards equivalent to the strictest European and (BRT) system known as the Metrobus to supplant US standards (Euro VI or US EPA 2010) as part of informal buses known as colectivos and reduce its voluntary inspection and maintenance program private vehicle use along major routes. Since then, (see Table 4). Nationwide implementation of the Metrobus has expanded to seven lines, serving TABLE 4: Timing of Mexico’s Emission Standards and Alignment with Standards from the United States and the European Union Heavy-duty engines Complete Vehicles Implementation NOM-044 Aligned standard NOM-044 Aligned standard 1A U.S. 2004 3A California LEV I Until June 30, 2019 2A Euro IV 4A Euro 4 January 1, 2019– 1 AA U. S. 2007 – – December 31, 2020 2 AA Euro V 4 AA Euro 5 1B U.S. 2010 3B U.S. 2010 From January 1, 2019 2B Euro VI 4B Euro 6 Source: The International Council on Clean Transportation, “Mexico Heavy-Duty Vehicle Emission Standards,” 22 February 2018, https://www.theicct.org/publications/mexico-heavy-duty-vehicle-emission-standards Norma Oficial Mexicana NOM-086-SEMARNAT-SENER-SCFI-2005, Especificaciones de los Combustibles Fósiles para la Protección Ambiental. 20 28 | CLEARING THE AIR: A TALE OF THREE CITIES 800,000 passengers per day, with about 180 million different driving restrictions (see Table 5). In 2016, passenger trips per year. Line 1 of the Metrobus was in response to an air quality emergency, the federal registered under the Clean Development Mechanism government established more stringent emission (CDM) in August 2011 for a crediting period standards for all vehicles in the MACA (NOM-EM- between 2012 and 2019.21 The rest of the lines were 167-SEMARNAT-2016) that, in turn, meant that registered as a separate CDM project in September only vehicles equipped with an on-board diagnostic 2012 for a crediting period between 2013 and 2022.22 system to monitor engine emissions were qualified to The Metrobus reported a reduction of 143,952 tons obtain a hologram that exempted them from driving of carbon dioxide equivalent in 2016.23 restrictions. The norm also updated the technological requirements of authorized verification centers in • Driving restrictions: Since 1989, Mexico City has the MCMA to be able to access data from vehicles’ enforced driving restrictions on a rotating daily on-board diagnostic system and create a centralized basis according to vehicle license plate numbers depository of data that could be audited by federal (locally known as Hoy No Circula). When imposed authorities and used to inform new policies. The in 1989, the restrictions applied to 2.3 million norm established remote detection test methods to vehicles, or 460,000 vehicles per day (Davis 2008). collect data from ostensibly polluting vehicles. Prior In response to these restrictions, an estimated 22 to the adoption of the norm, only private vehicles percent of drivers purchased a second vehicle, were subject to the vehicle verification program. frequently older and more polluting units that The new norm also included vehicles that provide contributed to worsening air quality (Eskeland public and private services regulated by the federal and Feyzioglu 1995). The program was reformed in or state governments. In 2017, the emergency norm 1997 to exempt new vehicles (those manufactured was substituted by permanent norm (NOM-167- after 1993) equipped with catalytic converters, SEMARNAT-2017). As of May 2020, governments of thereby accelerating vehicular renewal (Blackman the states that are part of CAMe have been working to et al. 2018). Authorities have revised the driving harmonize these rules and the associated inspection restrictions several times with the aim of increasing programs across jurisdictions. program efficiency. 8. Beyond measures to reduce vehicular pollution, the • Vehicle inspection program: Mexico’s vehicle MCMA has taken important steps to reduce pollution inspection program that verifies each vehicle’s from industries. Industries emit a smaller share of total emissions is a key complement to the driving emissions in the MCMA compared to area sources and restrictions program. Most vehicles must be inspected motor vehicles, largely due to the interventions supported semi-annually at an authorized verification center.24 by the air quality programs mentioned above, which led Based on the results of the inspection, vehicles receive to the closure of heavy industries, including the refinery a hologram with a number that is associated with (known as 18 de marzo), cement factories, and smelters. TABLE 5: Driving Restrictions in Mexico City Hologram Weekday Restrictions Saturday Restrictions 00 and 0 Exempted Exempted 1 Unable to drive one day between Monday and Friday Unable to drive two Saturdays per month 2 Unable to drive one day between Monday and Friday Unable to drive on Saturdays Source: Secretaría del Medio Ambiente. 2020. Aviso por el que se da a conocer el Programa de Verificación Vehicular Obligatorio para el primer semestre del año 2020. Gaceta Oficial Distrito Federal, January 2, 2020. https://www.sedema.cdmx.gob.mx/storage/app/media/f835e1f556a4143a44ccbc8a76b92dcf.pdf. 21 https://cdm.unfccc.int/Projects/DB/AENOR1309257514.77/view. 22 https://cdm.unfccc.int/Projects/DB/SQS1347027039.13/view. 23 https://www.metrobus.cdmx.gob.mx/dependencia/acerca-de/reduccionemisiones. 24 Electric, hybrid and brand new vehicles that meet low emissions criteria are exempted from the verification program for period of two and eight years, depending on their characteristics. CLEARING THE AIR: A TALE OF THREE CITIES | 29 In addition to the measures implemented in the 1990s, in the area expanded their air quality management efforts to convert facilities to use cleaner fuels such as strategy to include such measures as controlled early- natural gas or electricity continued under the 2002-2010 season burning in the surrounding forests to reduce and 2011-2020 plans. NOM-086 also established a limit fuel loads and prevent large-scale, destructive fires.26 on the quantity of sulfur in industrial fuels used in the Improved control of wildfires was one of the 14 measures MCMA to 500 PPM. In 2010, established the Program to announced in December 2019 to improve air quality in Reduce Emissions of Atmospheric Pollutants to promote the megalopolis.27 the development and use of cleaner technologies, equipment and processes to improve air quality in the 10. Despite the tremendous progress made since the MCMA. The Program exempted firms regulated by the 1980s, air pollution continues to be a challenge for federal government from environmental contingency the Mexico City Metropolitan Area, demonstrating restrictions if they invested in technological improvements the importance of sustained political and financial or energy efficiency measures that resulted in reduced commitment. Pollution concentrations have decreased emissions and/or if they implemented programs to steadily since 1990, but the mean annual concentrations reduce emissions from their transport fleet, to report of PM2.5 and PM10 have increased slightly since 2015 and reduce GHGs, or to improve their environmental and remain above the national standard. Ozone performance (SEMARNAT 2010). Mexico City and State concentrations have also fallen significantly since the of Mexico governments have adopted similar provisions 1990s but remain a perennial issue, with maximum for locally regulated industries, including in their most hourly concentrations frequently exceeding the national recent rules on environmental contingencies adopted in air quality standard. These conditions highlight the 2019 (PAMC 2019; SESM 2019) importance of continuously incorporating scientific evidence and lessons learned to strengthen air quality 9. Mexico City Metropolitan Area has also wrestled management, particularly as continuing trends such as with air pollution from household solid fuel use and urbanization and motorization remain associated with open burning. Studies in the 1990s and 2000s found increased pollution. that open biomass burning and household use of solid fuels for cooking contributed around 5-13 percent of 11. Civil society has played an indispensable role the ambient PM2.5 concentrations in Mexico City (Vega in ensuring that commitment to cleaner air in et al. 2010). Within the megalopolis (CAMe’s geographic the MCMA is sustained. A participatory approach, coverage), fuelwood and agricultural burning were the incorporating public opinion is required to establish source of 39 percent of PM10, 50 percent of PM2.5 and 60 the legitimacy of actions to improve air quality. Local percent of black carbon emissions in 2015 (PROAIRE of authorities and CAMe have used social media and other the Megalopolis 2017-2030). Cooking with solid fuels channels to disseminate hourly air quality information to also results in household air pollution, that mainly affects raise awareness, provide actionable recommendations to vulnerable groups, given that cooking with solid fuels groups at risk from air pollution, and elicit stakeholders’ is almost exclusively concentrated among the poorest perspectives. Since February 2020, authorities have households (INSP-UNICEF 2016). During the dry season implemented a new federal standard that requires the (November to April), forest fires and agricultural burning use of a new color-coded index, called the Air and Health can contribute an even larger share of PM pollution Index, to better communicate air quality and empower (Lei et al. 2013). Episodes of high PM2.5 concentrations individuals to take actions to reduce their exposure to air in the region were associated with the increased pollution (NOM-072-SEMARNAT-2019). Public access to frequency of wildfires during 2019, for example.25 data and the legitimacy it engenders for public entities, as Recognizing the problem of open burning, authorities well as an evidence-based approach, has allowed CAMe 25 Secretaría de Salud de la Ciudad de México (Secretariat of Health of Mexico City). 2019. “Sobre el impacto en la calidad del aire por los incendios en el Valle de México.” https://salud.cdmx.gob.mx/comunicacion/nota/12052019-sobre-el-impacto-en-la-calidad-del-aire-por-los-incendios-en-el-valle-de- mexico. 26 Breathelife2030.org, “Mexico’s Mega City Advances the Fight for Cleaner Air,” Climate and Clean Air Coalition, U.N. Environment, 13 August 2018, http://ccacoalition.org/en/news/mexicos-mega-city-advances-fight-cleaner-air. 27 Comisión Ambiental de la Megalópolis. 2019. Medidas para mejorar la calidad del aire en la Zona Metropolitana del Valle de México. https://www.gob. mx/cms/uploads/attachment/file/518260/Presentacio_n_final_de_medidas_ZMVM-20dic2019__2_.pdf. 30 | CLEARING THE AIR: A TALE OF THREE CITIES and its predecessor, CAM, to work at a technical level quality standards, its experience shows that it is possible across administrative, political, and sectoral boundaries. for cities to reduce air pollution significantly over a short Also, CAMe and CAM have been able to set long-term period of time. goals, while maintaining flexibility to respond to events on the ground, which has proven indispensable to address 13. Improvements in air quality in Beijing reflect air pollution, which is a complex and long-term problem. the culmination of nearly two decades of effort to reduce pollution in the capital city and surrounding region. Figure 9 provides a condensed timeline of air 3B. AIR QUALITY MANAGEMENT IN BEIJING pollution control measures in Beijing city leading up to AND THE GREATER JING-JIN-JI (JJJ) REGION 2013. Although efforts to control emissions from single point sources, such as coal-burning industrial facilities 12. Not so long ago, Beijing made the list of the most and power plants, began as early as the 1970s, it was polluted cities in the world; since 2013, however, it not until the late 1990s that the city first put in place a has seen remarkable progress (see Figure 8). Though comprehensive strategy to tackle pollution from a variety Beijing still has a way to go to meet China’s national air of sources. But then measures focused on the core urban FIGURE 8: Monthly PM2.5 concentrations in Beijing, February 2009 to October 2018 500 APPCAP introduced 400 Monthly PM2.5 (μg/m3) 300 200 100 0 12 15 17 18 09 09 10 10 11 11 12 13 13 14 14 15 16 16 17 18 n l n l n l n l n l n l n l n l n l n l Ju Ju Ju Ju Ju Ju Ju Ju Ju Ju Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Monthly 5th and 95th percentile, U.S. embassy station Monthly mean, U.S. embassy station Monthly 5th and 95th percentile, Beijing MEMC network Monthly mean, Beijing MEMC network Trend in monthly mean Trend in monthly 95th percentile Notes: “APPCAP” = Air Pollution Prevention and Control Action Plan; “Beijing MEMC” = Beijing Environmental Protection Bureau Municipal Environmental Monitoring Center. ” U.S. embassy data are from the reference-grade continuous air quality monitor installed at the U.S. embassy in Chaoyang District, Beijing. Beijing city data are averaged from the network of 37 stations for which PM2.5 data are available from December 2013 to October 2018. Because the U.S. embassy data are only for a single location and monitor, the data prior to December 2013 should be viewed with caution. Source: Beijing city data from Beijing Municipal Environmental Protection Monitoring Center; U.S. embassy data are from U.S. Department of State ” http://stateair.net/. Air Quality Monitoring Program, “StateAir, CLEARING THE AIR: A TALE OF THREE CITIES | 31 FIGURE 9: Timeline of air quality management actions in Beijing, 1998-2013 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 1998-2010 2000-2006 2005- 2007 2010- Power plants Dedusting Desulfurization Coal to gas Denitrification Natural gas thermal-power retrofit retrofit retrofit cogeneration centres 1998-2002 2003-2008 Renovation of coal-fired 2009-2012 Renovation of coal-fired Renovation of small boilers boilers (capacity below 14 MW) in boilers (capacity over 14 MW) in six in core areas core areas urban districts Coal-fired boilers 2006-2009 Centralized heating supply centers in rural districts/counties 1998-2002 Residential 2003-2013 - Conversion to electric heating and energy efficiency Replacement with electricity for heating retrofits in old single-story houses in core urban districts selected old one-storey houses Emission standards Light – duty 1999/1-2002/12 2003/1-2005/12 2005/12-2008/2 2008/3-2013/1 2013/2- gasoline vehicles China 1 China 2 China 3 China 4 Beijing 5 Heavy – duty 1999/1-2002/12 2003/1-2005/12 2005/12-2008/2 2008/7- 2013/2- diesel vehicles China I China II China III China IV1.2 Beijing V1 Improve 2003/12005/6 2005/6-2007/12 2008/1-2012/5 2012/5- gasoline quality Uncleaded Reducing <500 ppm <150 ppm <50 ppm <10 ppm gasoline sulfur since 1998 content Improve diesel 2003/12005/6 2005/6-2007/12 2001/72012/5 2012/5- quality <500 ppm <350 ppm <50 ppm <10 ppm I/M programmes and 2001-2003 2003- 2013- in-use inspections Transition Full implementation of ASM and lug down methods On Road RS3 Traffic restrictions on 2003-Yellow-labelled vehicles specific in-use fleets 2004-Freight trucks and motorcycles Clean energy and new 1999- 2009-New energy vehicles promoted by the 2012- energy vehicles CNG public buses “Ten City & Thousand Units” programme LNG buses 2008/10-Driving restriction on 2011- light-duty passengers cars License control Traffic management and economic measures 2009-Subsidized scrappage of 2011-expanded to yellow-labelled vehicles older vehicles Temporary measures for the Oddeven restrictions, ban on yellow-labelled vehicles, Olympic Games reducing use of government-owned vehicles Air quality monitoring Pre-2008 - 27 monitors in city, 2013 –35 stations with continuous and forecasting PM10 monitoring begins PM2.5, all data made public 2001 – public air quality 2012 – AQI forecasting under forecasting begins new national standards Source: Reproduced from UNEP (2016). 32 | CLEARING THE AIR: A TALE OF THREE CITIES districts of Beijing, gradually expanding to cover other year earlier, the Ministry of Environmental Protection parts of Beijing Municipality.28 They had limited success (MEP)29 had issued new ambient air quality standards because a large share of ambient PM2.5 pollution in and ordered cities in the JJJ and other key regions to begin Beijing originated from other areas in the Jing-Jin-Ji (JJJ) publicly monitoring PM2.5 in accordance with the new region (see Figure 11), which apart from the municipality standards (MEP 2012b).30 Then, in January 2013, Beijing of Beijing also includes the municipality of Tianjin, the and much of eastern and northern China experienced province of Hebei, and small parts of Henan, Shanxi, several weeks of severe pollution, with concentrations inner Mongolia, and Shandong. Since 2013, with the of PM2.5 reaching levels similar to those experienced in support of the national government and line ministries, Delhi during November 2017 (see Figure 10). Following Beijing and the other cities in the JJJ have worked in this episode, in September 2013, the State Council coordination to tackle sources of emissions across the issued the national-level Air Pollution Prevention and region, irrespective of administrative jurisdiction. Control and Action Plan (APPCAP) for 2013 to 2017. The APPCAP set a target for Beijing to reduce mean annual 14. The year 2013 in fact marked a shift in strategy concentrations of PM2.5 to around 60 µg/m3 in 2017 (that in the fight against pollution in Beijing and the JJJ is, about 33 percent), while calling for the rest of the region, signaled by the intervention of the State JJJ region to reduce mean annual PM2.5 by around 25 Council, China’s highest governmental body. A percent. FIGURE 10: Daily average PM2.5 concentrations during severe pollution episodes in Beijing (January 2013) and New Delhi (November 2017) 600 8 - 12 Jan, 569 μg/m3 8 Nov, 576 μg/m3 12 Nov, 499 μg/m3 24-hour mean PM2.5 (μg/m3) 29 Jan, 430 μg/m3 400 200 11 Oct 19 Nov 11 Feb 0 1 Jan 0 10 20 30 40 Day New Delhi, 2017 Beijing, 2013 Notes: Notes: data are for illustrative purposes only; measurements are obtained by different instruments and have not been calibrated or adjusted according to a common standard or reference conditions. Sources: hHourly monitoring data from U.S. Embassy in Beijing, U.S. Department of State Air Quality Monitoring Program, http://www.stateair.net/; daily monitoring data from the Central Pollution Control Board for four stations in central New Delhi (Mandir Marg, Shadipur, Punjabi Bagh, and ITO), http://www.cpcb.gov.in/CAAQM/mapPage/frmdelhi.aspx?stateID=6 28 The downtown area includes East City District, West City District, Haidian District, Chaoyang District, Fengtai District, and Shijingshan District. These, along with peri-urban districts and counties, make up Beijing Municipality 29 In 2018, the MEP went through a milestone institutional reform to be the Ministry of Ecology and Environment (MEE). 30 The MEP order applied to cities at the prefecture level and above, which correspond roughly with districts in India. CLEARING THE AIR: A TALE OF THREE CITIES | 33 15. More detailed city, provincial and regional plans accomplishments by Beijing city during the APPCAP were issued following the National Action Plan. period included: The same day the national APPCAP was issued, Beijing published its Beijing City 2013-2017 Clean Air Action < Reducing coal consumption from 23 million tons in Plan. The 2013-2017 plan built on the city’s existing 2012 to less than 5 million tons in 2017 through such plan for 2011-2015 and the longer-term air quality measures as shutting down the remaining four coal- management strategy that it had already developed for fired thermal power plants in the city, replacing coal- 2012-2020, establishing quantifiable targets and defining fired heating stoves in more than 300,000 households the responsibilities of 42 offices and agencies and 23 with gas or electric systems, and converting or enterprises in the city for meeting those targets.31 Hebei scrapping 39,000 commercial and industrial coal- and Tianjin also released their five-year provincial-level fired boilers; strategies for 2013-2017 the same day the APPCAP was issued. A few days later, the MEP, National Development < Promoting public transport, tightening emission and Reform Commission (NDRC), Ministry of standards for vehicles, controlling fleet numbers, and Industry and Information Technology (MIIT), Ministry upgrading high-emitters such as diesel trucks and of Finance (MoF), Ministry of Housing and Urban- buses by replacing them with clean-energy vehicles Rural Development (MoHURD), and National Energy or by requiring that diesel particulate filters be used; Administration (NEA) issued a joint five-year action plan for air pollution prevention and control for the entire < Controlling industrial pollution by shutting down JJJ airshed. or relocating 1,992 large industrial enterprises in heavy sectors such as steel, chemicals, and building 16. The measures in the 2013-2017 city, provincial materials and providing financing for retrofits and and regional plans for air pollution control in the JJJ emissions controls in others; targeted a variety of sources identified as major causes of high PM2.5 levels. Pollution in the JJJ region comes < Controlling windblown dust by expanding tree cover from a diverse set of sources, as illustrated for Beijing, and green space by about 74,000 hectares (equal to Tianjin, and Shijiazhuang (the capital of Hebei province) 4.5 percent of the area of the entire municipality); in Figure 11. The main local sources in the JJJ cities retrofitting trucks hauling construction waste to generally include coal use (by industry, power plants, make them airtight; and adopting the new road and households); industrial process emissions; vehicular cleaning technique of “vacuum, sweep, wash, and exhaust; and fugitive dust (from construction sites, road collect” to prevent fugitive dust. dust, and other sources). Although not shown explicitly in Figure 11, secondary PM2.5 formed from other pollutants 17. Actions to improve air quality in Beijing and also constitutes a large share of total ambient PM2.5 the greater JJJ region have extended well beyond concentrations, including about 39 percent of PM2.5 in municipal boundaries. In 2015, for example, the Hebei Tianjin (Shi et al. 2018) and about 45 percent in the provincial Agricultural Bureau launched its Action Plan industrial city of Xingtai in Hebei (Chen et al. 2017). for Controlling Agricultural Non-Point Source Pollution Coal use by industries, powerplants, and households, (2015-2018). Under the plan, the province set a goal motor vehicle exhaust, and fertilizer use all contribute to to reduce air emissions from crop stubble burning secondary PM2.5 in the region. A large share of pollution by reutilizing 95 percent of the 60 million tons of in the region’s cities also comes from regional sources agricultural residues generated annually.32 That same outside municipal boundaries, underscoring the need year, the province also introduced new regulations for an airshed-scale management approach. Some of the banning open burning. According to official statistics, 31 Luo Qianwen, “Beijing City Issues 5-Year Clean Air Action Plan” (in Chinese), Beijing Daily, 13 September 2013, http://www.gov.cn/fwxx/sh/2013-09/13/ content_2487793.htm. 32 Lv Xiaohong, “Hebei’s Three-Year Action Plan for Reducing Non-Point Source Agricultural Pollution” (in Chinese), Fenghuang News, 22 July 2015, http://hebei.ifeng.com/news/detail_2015_07/22/4136782_0.shtml; Ministry of Agriculture, “Special Report on Battle Against Non-Point Source Pollution from Agriculture (2016 Report No. 10)” (in Chinese), 26 December 2016, http://www.moa.gov.cn/ztzl/mywrfz/jbxx/201612/t20161226_5417353.htm; World Bank (2016). 34 | CLEARING THE AIR: A TALE OF THREE CITIES FIGURE 11: Sources of PM2.5 pollution in cities in the JJJ region a. Beijing city, 2012-2013 local pollution emissions 64%-72% coal combustion 14.1% 22.4% industrial production 14.3% motor vehicles 18.1% dust 31.1% region transmission 28%-36% other local pollution emissions b. Tianjin city, 2012-2013 local pollution emissions coal combustion 66%-78% 6% industrial production 27% 30% motor vehicles 17% dust region transmission 20% 22%-34% other local pollution emissions c. Shijiazhuang city, Hebei province, 2013-2014 local pollution emissions coal combustion 70%-77% 8.8% industrial production 28.5% 22.5% motor vehicles region transmission 15.0% dust 25.2% 23%-30% other local pollution emissions Source: figures reproduced from ICCS (2018) CLEARING THE AIR: A TALE OF THREE CITIES | 35 by 2017 the reutilization rate for agricultural residues resources to leverage an additional RMB 20 billion (US$ had reached 96 percent, with most being returned to 2.96 billion) in financing from the private sector. the field to enhance soil nutrient content or being used for animal feed. Still, provincial authorities report that 19. Financing from the central government and city enforcement continues to be a problem and burning for air pollution control in Beijing has supported a has continued in some areas.33 In addition, to tackle variety of incentive programs. Examples of incentive generation of secondary PM2.5, the local government programs include: in Hebei initiated a pilot program (under the World Bank supported Hebei AQM Program for Results (P4R) < Subsidies for end-of-pipe controls and boiler project) to increase Nitrogen Use Efficiency (NUE) on retrofits in power plants and factories: Since 2004, 6.2 million hectares of farmland that will reduce NH3 the national government has offered coal-fired power through balanced fertilizer application. This pilot is plants a price subsidy to offset the increased marginal expected to be further expanded to other parts of China. cost of installing and operating end-of-pipe controls.36 China is also: (i) piloting the use of covered tanks for Smaller industries have also received incentives for storage of animal manure to minimize NH3 emission installing control technologies and lower-emissions on large livestock farms; (ii) applying technologies that equipment. In 2017, businesses retrofitting small ensure that fertilizer is injected sufficiently into the soil boilers (steam capacity up to 20 tons) to reduce NOx on large farms for both crop and vegetable production; emissions could receive subsidies from the city of and (iii) developing new plasma reactor technology that RMB 96,000 to RMB 580,000 depending on the size transforms manure from liquid to solid form prior to of the boilers. District governments in the city were being used on farmland. expected to offer matching subsidies.37 18. The central government has provided dedicated < Cash for clunkers: By 2015, Beijing had already financial resources to support implementation of banned the worst-polluting vehicles from entering pollution control strategies in the JJJ region. All told, the city.38 Additionally, it offered cash incentives between 2013 and 2017, direct outlays by the central for owners with vehicles that had been registered government totaled RMB 52.8 billion (US$ 7.81 billion) for at least 6 years to scrap their older vehicles. The in special-purpose funds and RMB 10 billion (US$ 1.48 incentives ranged from RMB 3,500 (US$ 518) for billion) in budgetary resources to support air quality small passenger vehicles to RMB 21,500 (US$ 3,180) management in the JJJ provinces.34 In 2016, for example, for large buses. Vehicle owners had to dispose of Hebei province received earmarked resources of about their vehicles at least a year ahead of the mandatory RMB 4.01 billion (US$ 593 million) from the central retirement date, if applicable (e.g. for example, for government plus RMB 902 million (US$ 130 million) public buses), in order to receive the subsidy. The in transfers from Beijing and Tianjin cities, while city has also offered subsidies to encourage drivers to committing RMB 800 million (US$ 120 million) in purchase clean-energy vehicles. In 2017, for example, specially designated provincial resources to implement drivers could receive RMB 44,000 (US$ 6,509) in its air pollution control plan.35 The province used these national subsidies plus RMB 22,000 (US$ 3,254) in 33 Ministry of Agriculture, “Hebei Maintains Reutilization Rate for Agricultural Residue above 96%,” 5 June 2017, http://www.moa.gov.cn/ztzl/2017sxsc/ sxdt/201706/t20170605_5660829.htm; Sohu News, “Reutilization Rate for Agricultural Residues Reaches 96% This Year, Still a Structural Problem” (in Chinese), 4 December 2017, http://www.sohu.com/a/208280431_115479. 34 Data on central government support for air pollution control are from remarks by Liu Bingjiang, MEP Atmospheric Division Chief, GoC, “Responses from Ministry of Environmental Protection on Progress Made in Air Pollution Prevention and Control Work” (in Chinese), 27 February 2018, http:// www.gov.cn/xinwen/2018-02/27/content_5269486.htm. 35 Gao Zhili, Director of Hebei Province Financing Department, “Report on the Implementation of the 2016 Budget and Drafting of the 2017 Budget in Hebei Province” (in Chinese), http://www.mof.gov.cn/zhuantihuigu/2017ysbghb/201703/t20170306_2547723.html. 36 In 2013, power plants received electricity price subsidies from desulfurization, denitrification, and dust-removal equipment of RMB 0.015/kWh, RMB 0.01/kWh, and RMB 0.002/kWh, respectively. Christopher James, “China’s Power Sector and Air Quality Reforms: Global Lessons on Getting Institutional Responsibilities Right,” 7 November 2017, https://www.raponline.org/knowledge-center/chinas-power-sector-air-quality-reforms-global-lessons-getting- institutional-responsibilities-right/?sf_action=get_data&sf_data=results&_sft_region=china&_sft_language=english. 37 Beijing, Sanhui Heating and Environmental Technology Development Company, Limited, “Beijing City Subsidy Policies for Low-Nitrogen Retrofits of Boilers” (in Chinese), http://www.bjshnh.com/gaizao/91.html. 38 Called “yellow-label” vehicles for the color of their emissions stickers, these included gasoline vehicles that failed to meet the China I standard and diesel vehicles that could not meet the China III standard. 36 | CLEARING THE AIR: A TALE OF THREE CITIES subsidies from Beijing city for buying all-electric issues such as fuel standards, energy supply, and public vehicles.39 transportation.41 In July 2018, the State Council decided to further elevate the administrative status of the < Household subsidies for clean heating: By October coordination group, with the Minister of Ecology and 2017, Beijing had imposed a total ban on coal use Environment (MEE), Beijing City Governor, Tianjin City by households for cooking and heating within city Governor, and Hebei Provincial Governor now acting limits, including in peri-urban villages. To help as its vice-chairs. Other members now include vice- families make the switch, the city heavily subsidized ministers from the NDRC, MIIT, Ministry of Finance, the purchase of new boilers and stoves and lowered Public Security Administration, Ministry of Housing the price of energy for heating. Households in rural and Urban Development, Ministry of Transportation, areas switching to electric heating, for example, Ministry of Agriculture, Chinese Meteorological could receive RMB 24,000 (US$ 3,550) in subsidies Administration, and other agencies. The Secretariat is for a new heating system, reducing the upfront cost expected to sit within the national-level MEE (formerly per household to around RMB 3,000 (US$ 444). MEP).42 Additionally, they enjoyed a lower tariff rate for electricity use during winter nights, reducing the 21. Local government officials and state-owned- average winter heating cost per household to around enterprise managers were held accountable for RMB 1,661 (US$ 246). By comparison, the average implementing air quality management plans by winter heating cost for a household burning coal integrating air quality targets and measures into would have been around RMB 5,000 (US$ 740).40 official performance reviews. In 2014, the State Council issued rules for evaluating the performance of 20. Cross-jurisdictional coordination on air pollution government officials in carrying out the responsibilities control in the JJJ region has been led by the State defined in the local-level air pollution prevention and Council and the JJJ Regional Air Quality Prevention control plans.43 Per the rules, officials failing to meet the and Control Coordination Group, with high-level 2017 targets would be ordered to appear before the State participation. Provincial and city-level governments Council, and approvals for new development projects continued to be the primary implementers for air in their area would be suspended. The rules also made quality management in the JJJ region. At the same time, the achievement of air quality targets a more integral the JJJ Regional Air Quality Prevention and Control part of official cadre performance reviews. China’s Coordination Group was established in 2013 to ensure government first started to embed environmental targets cross-jurisdictional coordination. Originally housed in the system for evaluating cadres in the early 2000s. The in the Beijing Environmental Protection Bureau, the annual reviews can determine bonuses and an official’s group included representatives from seven provinces prospects. Although far from perfect (see Kostka, 2015), and municipalities and eight ministries under the State the reviews have become an influential lever in shaping Council, with the Vice Premier serving as its head. official behavior (Wang, 2013). Officials that consistently Mirroring the regional group, multi-agency leadership fail to meet targets may be demoted, reassigned to a groups were also established at the provincial and local less prestigious locality, or possibly expelled (Eaton and levels of government. The ministries and provinces in Kostka, 2014). In the closing months of the APPCAP , the regional coordination group formulated annual the MEP announced that it had disciplined 12,000 implementation plans for air quality management in officials.44 The government has taken continuous efforts the JJJ region and set policies for cross-jurisdictional to refine policies. For example, to enhance environmental 39 Subsidy amounts are for an all-electric vehicle with a range of greater than 250 km. Beijing People’s Government, “Beijing City Subsidy for New Energy Vehicles Set at 50% of National Subsidy, Maximum of RMB 44,000” (in Chinese), 10 February 2017, http://www.beijing.gov.cn/bmfw/zxts/t1468354.htm. 40 Xinhua, “Beijing Resolutely Switches from Coal to Electric Heating” (in Chinese), 22 April 2017, http://www.xinhuanet.com/politics/2017- 04/22/c_1120854758.htm; He Cong, Jia Yong, Xi Bo, Wan Jintao, “Another Heating Season Arrives, How Can the North Have Heat While Also Having Good Air?” (in Chinese), People’s Daily, 15 November 2017, http://www.bjd.com.cn/sd/hcr/201711/15/t20171115_11074943.html. 41 Beijing Environmental Protection Bureau Press Office, “Air Pollution Prevention and Control Work in the Expanded Jing-Jin-Ji Region” (in Chinese), 24 July 2014, http://bjepb.gov.cn/bjhrb/xxgk/jgzn/jgsz/jjgjgszjzz/xcjyc/xwfb/607628/index.html. 42 General Office of State Council, Notice of General Office of the State Council Regarding the Formation of the Leading Group for Air Pollution Prevention and Control in the Extended Jing-Jin-Ji Region, 11 July 2018, http://zhengce.beijing.gov.cn/library/192/33/50/438650/1558190/index.html. 43 State Council, “Notice of the General Office of the State Council on the Issuances of Rules for Evaluation of Implementation of the Air Pollution Prevention and Control Plan” (in Chinese), 30 April 2014, http://www.gov.cn/zhengce/content/2014-05/27/content_8830.htm. CLEARING THE AIR: A TALE OF THREE CITIES | 37 performance accountability at the city level, provinces closely with their own three-year strategies. As air quality such as Hebei and Shanxi introduced a transparent fiscal continues to improve, many of the cheapest no-regrets budget reward and penalty scheme to take account of solutions have already been taken, and the challenge municipal Air Quality Index (AQI), PM2.5, PM10 and SO2 for the next round of air pollution management will be concentrations levels using a formula. for the governments in the JJJ to analysis and identify measures that deliver the greatest reduction in exposure 22. Regulatory changes and greater openness in the at the least cost to the economy. Furthermore, innovative public dissemination of environmental data have also financing mechanisms and private sector investments helped improve government accountability for air will need to be mobilized to sustainably finance air quality. Rules promulgated by the MEP have required quality management for the long term. that they make their environmental monitoring data public, and near-real-time data on air quality is now 3C. AIR QUALITY MANAGEMENT IN DELHI available for most cities and stations. The government 25. There is a long history of tackling air pollution in has also required continuous emissions monitors to be India, with supporting legislation. The main legislation installed in large industrial facilities and for the data that governs air pollution management in India is the from these facilities to be made public. Air (Prevention and Control of Pollution) Act (“Air Act”), first issued in 1981 and amended in 1987.47 The 23. The JJJ region made tremendous progress in Air Act empowers the Ministry of Environment, Forest, reducing PM2.5 pollution during the APPCAP period, and Climate Change (MoEFCC), and its subordinate finally bending the curve on rising concentrations. institutions, Central Pollution Control Board (CPCB) Average PM2.5 concentrations in Beijing exhibit a distinct and the state pollution control boards (SPCBs), to drop beginning around September 2013, as the city perform a range of functions to help prevent, control, and reduced its mean annual PM2.5 from around 90 µg/m3 to abate air pollution in the country. CPCB, for example, 58 µg/m3 in 2017 (see Figure 8). Severe pollution episodes has the responsibility to prescribe air quality standards, have also abated. Mean annual PM2.5 in the larger JJJ the first set of which were issued in 1984 and have been region dropped by nearly 40 percent, far exceeding since been expanded and strengthened. Additionally, the the APPCAP target of 25 percent.45 Some cities in the Air Act requires CPCB and SPCBs to develop nationwide JJJ region achieved even larger declines, including the programs and state-level plans, respectively, to help industrial center of Xingtai in Hebei province, which had achieve ambient air quality standards. Program and plan some of the worst pollution in the country at the start implementations are left to state governments, however. of the APPCAP. Mean annual PM2.5 in Xingtai declined from 160 µg/m3 in 2013 to just below 80 µg/m3 in 2017.46 26. Pollution control boards also have powers to prescribe and enforce emission standards for 24. Much remains to be done before the cities will stationary and mobile source of air pollution, in reach compliance with the national standards, coordination with other government agencies. For however. By early 2018, average levels of PM2.5 in the stationary sources of pollution – industrial plants and JJJ remained about two times higher than the national thermal power plants, for example -- India relies on standard and seven times higher than the guideline value an emission-standards-and-permit system. Emission recommended by the WHO. To continue the momentum standards for different categories of industry (including of the APPCAP , in August 2018 the State Council issued power plants) have been established, and facilities the Three-Year Action Plan to Win the Battle for Clean must obtain permits to establish and operate, that, in Skies in 2018 to 2020. Beijing, Hebei, Tianjin, and turn, allows regulators to enforce the set standards (see other local governments in the JJJ region have followed 44 Nectar Gan, “12,000 Officials Disciplined and 18,000 Companies Punished in China’s Sweeping Crackdown Against Pollution,” South China Morning Post, 2 September 2017, https://www.scmp.com/news/china/policies-politics/article/2109342/top-level-china-pollution-inspections-wrapping. 45 By some estimates, better-than-usual atmospheric circulation and other favorable weather conditions contributed about 30 percent to the reduction in mean PM2.5 in 2017 compared to 2013. Li He, “Air Quality Takes a Turn for the Better in the Jing-Jin-Ji Region – Experts Credit More Focused, Targeted Pollution Control Efforts” (in Chinese), Xinhua, 28 February 2017, http://www.xinhuanet.com/energy/2018-02/28/c_1122463587.htm. 46 Hebei News, “PM2.5 Concentrations Drop 8%! Target Fixed for Xingtai to Improve Air Quality in 2018” (in Chinese), 27 April 2018, http://hebei.hebnews. cn/2018-04/27/content_6863379.htm. 47 Other Central Acts (Water, Environment Protection, Motor Vehicles, and Public Liability) also have provisions to regulate air quality (CSE 2016). 38 | CLEARING THE AIR: A TALE OF THREE CITIES Box 3). Regulators also monitor emission levels and emission standards, MoPNG fuel quality standards, and have the authority to shut down or disconnect water or state police and state transport departments enforce I and power to non-compliant units. Fines and imprisonment M programs. In critically polluted areas (including non- can be pursued through courts. For mobile source of attainment cities), pollution boards can also prohibit pollution, vehicles, as in much of the world, India relies the use of certain fuels and burning of any materials on emission standards for newly manufactured vehicles deemed to cause air pollution and mandate the use of along with fuel quality standards. Vehicle inspection certain pollution control equipment. For example, use and maintenance (I and M) programs are used for in-use of pet coke and furnace oil has been restricted in Delhi vehicles. Central Motor Vehicles Rules of 1989 grants and the National Capital Region (NCR), and brick kilns the Ministry of Road Transport and Highways (MoRTH) are required to use prescribed technologies to reduce air the responsibility to establish vehicle specifications. pollution in this area. This provision also allows SPCBs Therefore, standards for vehicles and I and M program to restrict trash burning in non-attainment areas. are developed jointly by MoEFCC and MoRTH. Ministry of Petroleum and Natural Gas (MoPNG) 27. Based on this legislation, but driven by public and the Bureau of Indian Standards are in charge of interest litigation and public pressure, Delhi began setting standards for quality of gasoline and diesel in to tackle air pollution starting in the mid-1980s. Air the country. Furthermore, MoRTH enforces vehicular pollution started to deteriorate in Delhi beginning in BOX 3: Controlling Air Pollution from Coal-Fired Power in India India has made great strides in extending access to electricity to its people. In 1990, only 43 percent of the population had some form of access to electricity. By 2016, that number had risen to nearly 85 percent.48 Still, about 25 million rural households continue to be unconnected,49 and per capita consumption of electricity in India in 2016 was less than a quarter that of the other BRICS countries, at 916 kWh compared to 4,262 kWh. The power sector in India has ample room to grow, and demand is expected only to rise. Coal has largely fueled India’s expanding access to electricity and will continue to do so in the coming decades. By the end of 2018, India’s power utilities had 197 gigawatts (GW) of installed coal-fired capacity, out of a total generating capacity of 347 GW, as the share of coal-fired generation in India’s total electricity supply reached 78 percent (see Figure B3.1). Separately, industries operated another 57 GW of captive thermal generating capacity FIGURE B3.1: Share of coal-fired power in total electricity generation in India, 1980-2018 85 Electricity generated from coal 80 75 70 (% total) 65 60 55 50 45 40 80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 10 12 14 16 18 20 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 Notes: The data include generation by utilities and exclude captive generation by industries. Sources: 1980-2014 data from IEA (2018b); 2015-18 data from India’s Central Electricity Authority (CEA) and Indiastat World Bank, “Access to Electricity (% of population)” (EG.ELC.ACCS.ZS), World Development Indicators database. 48 Bhasker Tripathi, “15 Million Indian Households Have Meters But No Electricity,” Bloomberg, 1 November 2018, https://www.bloombergquint.com/ 49 global-economics/15-million-indian-households-have-meters-but-no-electricity#gs.tUw7ok1D. CLEARING THE AIR: A TALE OF THREE CITIES | 39 for their own use. The International Energy Agency projects that, by 2040, the share of coal in total generation will drop to just below half, as India’s electricity mix becomes more diversified. Still, coal will continue to be the single largest fuel source, and, in absolute terms, coal-fired generation will be about twice what it was in 2017.50 Although expanding access to electricity has improved the lives of hundreds of millions of people, the growing reliance on coal-fired generation has also come at a cost. Coal-fired power is a major source of air pollutant emissions. Amid growing public concern over pollution, in 2015 the Ministry of Environment, Forest and Climate Change (MoEFCC) introduced new standards for emissions from thermal power plants, which were amended in 2018. The new standards restricted SO2, NOx, and mercury (Hg) emissions for the first time and tightened the existing limits on PM emissions. The new standards were comparable to those in other countries although above the limits currently achievable with the best available technologies (see Table B3.1). The MoEFCC set a deadline for power plants to comply with the standards by the end of 2017. The deadline was later extended to 2019 for Delhi and to 2022 for all other states and UTs. India’s Supreme Court has called for the government to implement the new standards by 2021 for 57 generating units located in India’s most densely populated areas.51 As of mid-2018, only 15 out of 650 units had installed emissions controls capable of meeting the new standards for SO2, and 273 units were in non-compliance with the PM standards. While some of the non-compliant units are older and will be retired, more than 410 units will need to be retrofitted with SO2 controls and over 230 units will need to have their PM controls upgraded to meet the 2022 deadline (CEA 2018). A range of viable technologies exist to control SO2, NOx, and PM before, during, and after combustion Pre- combustion controls include coal washing, which reduces the content ash and other impurities and improves the heating value of coal. During-combustion controls include optimizing the temperature, burn time, and boiler load and injecting sorbents such as limestone into the ame zone of the boiler. Post-combustion controls TABLE B3.1: Power Plant Emissions Standards in Various Countries Existing plants New plants NOx SO2 PM2.5 NOx SO2 PM2.5 China (2014) 100 50-200 20-30 50 35 10 India (2015) 300-600 200-600 50-100 100 100 30 Japan (2015) 410 100 200 200 50 South Africa 1,100 3,500 100 750 500 50 (2012) USA (2014) 135 185 18.5 95.3 136 12.3 165-220 (daily) 165-220 (daily) 11-20 (daily) 125 (daily) 110 (daily) 10 (daily) EU (2017) 65-85 (annual) 130-180 (annual) 8-12 (annual) 85 (annual) 75 (annual) 5 (annual) Notes: Existing plants refer to plants in operation as of when the respective standards were issued; new plants are plants that enter operation after the respective dates that the standards were issued or entered into force; the date of issue varies by region/country; the ranges shown in the table depend on the size, type, age, and location of the plant as per the standards for the respective countries/regions shown in the table; averaging or sampling periods vary; standards for the United States and Japan have been converted into units of µg per Nm3 by Zhu and Wang (2014); standards for China shown in the table are for the 11 provinces in the eastern region; standards for the EU shown in the table apply to a plant with a generating capacity of 300 MW to 1 GW and are effective as of 2021; all other standards are currently in effect. Sources: Zhang (2016) and EU (2017) Projections are for the “New Policies Scenario” in IEA’s World Energy Outlook 2018 (IEA 2018a). 50 Densely populated areas include areas with more than 400 people per square kilometer. Manka Behl, “Supreme Court to 57 Power Plants: Clean Up Act 51 within 28 Months,” TNN, 11 September 2018, https://timesofindia.indiatimes.com/city/nagpur/supreme-court-to-57-power-plants-clean-up-act-within-28- months/articleshow/65761232.cms. 40 | CLEARING THE AIR: A TALE OF THREE CITIES include end-of-pipe technologies to capture or remove pollutants from flue gas. Examples of new technologies to emerge in recent years include f lue gas conditioning, enhanced ESP , better fabric filters, agglomeration (binding small particles together to make them easier to capture), and hybrid systems (Zhang, 2016). In parallel with emissions controls, installing continuous emissions monitoring (CEM) equipment and making CEM data publicly available can improve compliance by ensuring that companies operate within standards and that the most egregious violators are held accountable. The MoEFCC has prescribed specific pollution control equipment for power plants to meet the new standards. New PM emission standards are to be met through the augmentation of ESP , f lue gas desulphurization to control SO2 is recommended as are post combustion measures like selective catalyst reduction and selective non-catalyst reduction systems for more stringent NOx standards. On its part, the Ministry of Power has taken measures to remove existing hurdles in the implementation of the new emission standards. The Central Electricity Authority, for example, has issued standard technical specifications for wet limestone-based flue gas desulphurization for a typical 500 MW unit which can readily be used by the utilities. Retrofitting thermal power plants with emissions control technologies can be technically challenging and expensive. Power producers will need to recoup the costs of putting in place control measures by raising tariffs. Ultimately, accelerating the retirement of older plants while promoting more renewable energy and energy efficiency may prove more cost-effective, allowing India to reduce harmful emissions from the power sector and avoid the lock-in effects that may also increase climate mitigation costs down the road.52 As part of its commitment to global climate change mitigation efforts, the Government of India has set an ambitious target of meeting energy demand through 175 GW of renewable energy. the 1980s. In 1985, public interest lawyer, M.C. Mehta, government to enforce the phaseout of leaded gasoline, asked the Supreme Court of India to direct government to introduce pre-mix fuels for two-stroke engines, agencies and departments to implement the Air Act in relocate polluting industries, and implement scrapping Delhi. The Supreme Court has been granted the authority of commercial vehicles in use for 15 years or more. The to hear cases that infringe on a citizen’s fundamental Delhi government and MoEFCC developed their first rights that are guaranteed under Article 21 of the Indian comprehensive plans for controlling pollution in 1996 Constitution. Right to a clean environment is one such and 1997, respectively, on the Court’s urging. The Delhi fundamental right. Spurred by the case, the Supreme government plan called for a Mass Rapid Transport Court began to push not only the Delhi government, System, a highway bypass around Delhi, improved but also the Central government, to take measures to vehicular technology, higher-quality fuels, increased use improve Delhi’s air quality. of CNG, restrictions on excessively polluting vehicles, urban greening, and a program of public awareness. The 28. Over the course of the next decade and a half, a MoEFCC plan was similar and focused on transport. number of measures were implemented based on The Court subsequently asked for a committee to be the push from the Supreme Court and pressure from established to monitor the progress on implementation the public, that brought some temporary relief from and to suggest other policies to control pollution. poor air quality. In the 90s, emissions from transport This committee was called the Environment Pollution were understood to contribute 60-70 percent of the (Prevention and Control) Authority (EPCA). Soon after, total, with the rest coming from industry and residential based on a report from EPCA, the Court mandated that sources; sources were understood to be located within all commercial vehicles switch to CNG by 2001. This the city boundaries. As a result, most measures to reduce order initiated one of the most ambitious vehicular- air pollution in Delhi focused on transportation (see fuel-conversion programs globally, and one that was Table 6). Between 1994 and 1998, the Court pushed the implemented in less than two years by December 2002. 52 Authors’ email correspondence with Gailius Draugelis, World Bank Lead Energy Specialist, February 2019. CLEARING THE AIR: A TALE OF THREE CITIES | 41 With coordinated action from a range of stakeholders 30. Delhi’s air pollution challenge has come to be -- suppliers of CNG, bus manufactures, public vehicle recognized as multi-sectoral and multi-jurisdictional. operators -- about 10,000 buses converted from diesel to Over the last few years, three different source CNG, and approximately 20,000 taxis and 50,000 three- apportionment studies (Sharma et al. 2016, Amman wheelers from petrol to CNG, bringing some brief relief et al. 2016, and ARAI-TERI 2018) have estimated the to Delhi’s citizens (see Figure 12). contribution of different sectors to seasonal (winter and summer) and mean annual ambient concentrations of 29. Air quality, however, steadily declined in the 2000s, PM2.5 in Delhi and the NCR. Findings from the latest of building up to a crisis in the late 2010s. As shown in these studies --- ARAI-TERI 2018 -- are illustrated in Figure Figure 12c, mean annual PM10 concentrations in Delhi 13. These studies show that there are multiple dominant have increased year after year from the early 2000s. The sources of pollution which require a multi-source strategy winter of 2016 brought very poor air quality to the city, for pollution management in Delhi. Though there are with PM2.5 concentrations of up to 759 µg/m3, 12.7 times differences, all three studies rank secondary inorganic higher than the national standard (EPCA-CSE 2018), aerosols (formed by reactions involving primarily NH3 for example. Delhi has consistently featured on top in from agriculture mixed with NOx and SO2); biomass WHO’s list of most polluted cities in the world during (household cookstoves); waste burning; and vehicular this period. emissions (including pollution from diesel vehicles, TABLE 6: Timeline of key measures to tackle air pollution in Delhi in the 1990s and early 2000s Date Measure 1985 Mr. Mehta petitions the Supreme Court (SC) to direct various government departments to enact the Air Act in Delhi. 1986 SC directs the Delhi Government to specify what steps have been taken to reduce air pollution. First set of vehicular exhaust emission standards are set (for in-use vehicles) and committee established to recommend 1990 vehicular emission standards (for new vehicles); Central Government approves Delhi Second Master Plan which calls for polluting industries to be relocated by 1993. Early 1990 MoEFCC proposes that all vehicles have compulsory catalytic convertors by 1995. SC asks MoEFCC to establish statutory committee (under provision of Environment Protection Act (1986)) to devise 1991 policies to reduce air pollution. The Saikia Committee is established. 1992 The Saikia Committee recommends phase-out of unleaded gasoline in Delhi and use of CNG as an alternative fuel. 1993 MoEFCC established vehicular emission standards to be met by 1996 and 2000. The Motor Vehicles Act is amended to encourage use of alternative fuels; SC mandates phase out of unleaded fuel in 1994 Delhi by 1995 and gradually reduction in the sulfur content of diesel by 1998. These deadlines are met. The Delhi Government announces subsidy for 2-3 wheelers to install catalytic convertors and MoRTH bans 1995 registration of 4 wheelers without catalytic convertors in Delhi. These initiatives are successful. With further deteriorations in air quality, SC orders Delhi Government to prepare a comprehensive air quality 1996 management plan. On orders from the Supreme Court, Delhi Government relocates highly polluting industries from Delhi having 1996-97 missed the 1993 deadline. Delhi Government and MoEFCC present first comprehensive actions plans for air quality management. Based on its Action Plan, Delhi Government develops policy to phase- out old vehicles starting in 1998 to 2000 and 1997 encourage the use of CNG. SC orders MoEFCC to establish the third in a series of committees to monitor implementation of action plans and devise new policies. This committee is called Environment Pollution (Prevention and Control) Authority (EPCA) 1998 and is still functioning. Based on recommendations from EPCA, SC orders that all commercial passenger vehicles -- autos, taxis, and buses -- switch to CNG by 2000. On orders from the Supreme Court, Delhi Government implements policy to phase out vehicles older than 15 years 1999 having missed previous deadline of 1998. Incentive scheme -- exemption from 8% sales tax and subsidy of 4% on the interest rate of the loan -- is used to incentivize conversion. 2002 SC's 1998 CNG conversion order is implemented. 42 | CLEARING THE AIR: A TALE OF THREE CITIES FIGURE 12: Monitored concentrations of pollutants in Delhi compared to national standards,1990-2018 a. Mean annual SO2 concentrations in Delhi NCR, 1990-2018 NAMP mean, 1990-2017 (2-7 stations) NAMP max annual mean CAAQM mean, 2017-2018 (7 stations) CAAQM max annual mean 60 National standard (2009) 50 Mean annual SO2 (μg/m3) 40 30 20 10 0 1990 1995 2000 2005 2010 2015 b. Mean annual NO2 concentrations in Delhi NCR, 1990-2018 140 130 120 110 100 Mean annual NO2 (μg/m3) 90 80 70 60 50 NAMP mean, 1990-2017 40 (2-7 stations) WHO guideline and national standard 30 NAMP max annual mean 20 10 CAAQM mean, 2016-2018 (9-10 stations) 0 CAAQM max annual 1990 1995 2000 2005 2010 2015 mean CLEARING THE AIR: A TALE OF THREE CITIES | 43 c. Mean annual PM10 concentrations in Delhi NCR, 2003-2018 500 450 400 350 Mean annual PM10 (μg/m3) 300 250 200 NAMP mean, 2003-2017 (3-8 stations) 150 NAMP max annual mean 100 CAAQM mean, 2016-2018 50 National standard (5-6 stations) WHO guideline 0 CAAQM max annual mean 2003 2005 2007 2009 2011 2013 2015 2017 Notes: For O3, CO, NO2, and SO2, data for 1990 to 2017 are from the manual monitoring stations in the National Ambient Air Quality Monitoring Programme (NAMP) network; for PM10, data for 2003 to 2015 are from the NAMP network; for all other years, data are from the automated stations in the Continuous Ambient Air Quality Monitoring (CAAQM) network as 2018 data were not available from the manual NAMP stations. Source: India Central Pollution Control Board, available at https://app.cpcbccr.com/ccr/#/caaqm-dashboard-all/caaqm-landing. road dust, tire wear, and brakes) as the most important members who have served longer term on EPCA include sources of pollution in the winter months. Confirming officers from Government of Delhi’s Environment these findings, Cusworth et al. (2018) find that smoke Department, Transport Department, Jal (Water) Board, from agricultural burning in upwind states contributed officials from Delhi’s four municipal corporations, and about 21-72 percent of ambient PM2.5 in Delhi during Delhi police. EPCA has continued to play a role in early November when the NCR frequently experiences advising the Supreme Court in matters pertaining to air severe pollution episodes. There is less agreement pollution management and to ensure implementation of among the studies on pollution sources in the summer air quality standards and related Court orders in Delhi months, though dust and construction are likely the and the NCR (MoEFCC 2018). In consultation with the largest contributors. Regional sources of air pollution Central government and relevant state governments, also contribute a large share of overall ambient PM2.5 EPCA developed a comprehensive action plan that concentrations, with as much as 60 percent of Delhi’s provides timelines for different responsible agencies to pollution comes from neighboring states (Amman implement medium and long-term measures to tackle et al. 2016). air pollution in Delhi NCR. This plan was subsequently notified by CPCB in 2018 per directions of the Court. 31. Delhi has once again taken on the challenge of The plan includes measures of reduce emissions from deteriorating air quality with EPCA continuing vehicles, power plants, industry, construction, waste to play an important role. The tenure of EPCA has burning, and solid fuels in domestic and hotel sectors. In been extended time and again, and most recently EPCA addition, an emergency action plan – Graded Response was reconstituted in October 2018 to include a larger Action Plan -- developed also by EPCA was notified number of representatives from civil society. Other in 2017 and becomes effective during high pollution 44 | CLEARING THE AIR: A TALE OF THREE CITIES FIGURE 13: Sources of PM2.5 pollution in Delhi NCR region, 2017-18 a. During winter months b. During summer months Delhi/NCR, 2017-2018 (winter) Delhi/NCR, 2017-18 (summer) motor vehicles dust biomass motor vehicles dust biomass industry other secondary industry other secondary Source: ARAI-TERI (2018) episodes initiating costly, short-term measures to bring 2017. New emissions standards for eighteen categories down pollution (EPCA 2019). For example, following of industries were issued in 2018 to reduce emissions the spike in air pollution on November 7, 2017, when of NOx and SOx in the NCR but will also contribute PM2.5 concentration in Delhi was 504 µg/m3, EPCA to better air quality nationally. The Badarpur coal- issued directives to state governments in the Delhi NCR based power plants that supplied power to Delhi to close polluting industries, continue ban on generator for over four decades was permanently shut in 2018. sets, stop use of unapproved fuels and coal and firewood Additionally, brick kilns in the NCR have been in eateries, among other measures to reduce dependence ordered to move to zig-zag technology, which will on private vehicles. substantially reduce pollution. 32. A number of measures have been implemented < To tackle agriculture residue burning: A National to improve air quality in Delhi, some at the national Policy for Management of Crop Residues was level and others specifically in the NCR. announced in 2014 and in 2018 the Cabinet Committee on Economic Affairs gave approval for < To tackle vehicular pollution: India has progressively the promotion of Agricultural Mechanization for tightened vehicular emission and fuel quality in-situ Management of Crop Residue in the states of standards and has announced plans to jump from Punjab, Haryana and Uttar Pradesh, and in National BS-IV (equivalent to ES-IV) to BS-VI (equivalent to Capital Territory of Delhi, in efforts to reduce large- ES-VI) from 2020, an unprecedented advancement of scale burning of crop residue from paddy crop in emission standards. Moreover, BS-VI (equivalent for October-November and wheat in April. ES-VI) grade fuel was made available in the National Capital Territory two years ahead of schedule. 33. Delhi, and other regions in the country, will also benefit from a number of other national initiatives < To tackle industrial and power plant emissions: to tackle air pollution, and others that though not Highly polluting fuels used in industry – petroleum initiated for air pollution concerns will nonetheless coke and furnace oil – were prohibited in NCR states help reduce air pollution as a co-benefit. In 2015, for as per the directions issued by CPCB in November example, MOEFCC introduced new emission standards CLEARING THE AIR: A TALE OF THREE CITIES | 45 for power plants, which, once enforced, will bring benefits these and other initiatives together with a concentrated in terms of reduction in PM2.5 concentrations (see Box 3). focus on air pollution management, in 2019, MoEFCC In addition, seventeen categories of polluting industries also launched the National Clean Air Programme have been directed to install Online Continuous (NCAP). The NCAP has set a time-bound goal for Emissions/Effluent Monitoring Systems, in an effort improving air quality across the country to help around to improve enforcement. Rules for construction and 100+ cities, where air pollution standards are currently demolition waste management were issued in 2016 and not being met, meet standards over time. The NCAP five waste management rules – for solid waste, hazardous provides cities an overall framework for developing air waste, plastic waste, biomedical waste, and e-waste quality management plans and guidance on interventions -- revised. The National Solar Mission and National across a range of sectors (power plants, transport, industry, Mission on Energy Efficiency that were initiated to agriculture, residential, waste management, and road and achieve India’s ambitious climate targets will also help construction dust management) to reduce air pollution. shift the energy mix towards renewables and away from The NCAP will also support the development of a thermal power plants, yielding air quality improvements national air quality monitoring network, dissemination as a co-benefit. India is also promoting electric vehicles53 of data, and involvement of the public in air quality and has initiated plans to develop hydrogen as a transport management planning and implementation. More fuel for the future. Programs such as the Pradhan Mantri recently, the Fifteenth Finance Commission of India, the Ujjwala Yojana that has expanded access to clean cooking constitutionally-mandated body established every five fuels, most notably LPG (see Box 4), reducing the reliance years to recommend sharing of tax revenues between the on residential biomass burning, is essential to efforts federal and the state governments, recommended a US$ to reduce air pollution. Ministry of Agriculture is also 500 million grant for one year to reward large cities for promoting programs to reduce the use of fertilizers such improvements in air quality. This performance-based as Soil Health Card Scheme, which provide soil fertility grant is a powerful, innovative mechanism to incentivize status and recommendations on per crop fertilizer dose, cities to act. and Integrated Nutrient Management practices. Bringing BOX 4: Pradhan Mantri Ujjwala Yojana Expands Clean Cooking in India The incomplete combustion of solid fuels such as wood, charcoal, dried dung, and crop residues in household cookstoves has long been a major source of both indoor and outdoor concentrations of air pollution in India. In 2000, only about 20 percent of households had access to clean fuels and stoves for cooking. Since then, India has made tremendous progress in expanding household access to clean fuels and technologies for cooking. By around 2017, the proportion of households with access to clean cooking had doubled. Successive initiatives launched by the central government over the past five years have accelerated the transition toward clean fuels and technologies for cooking, most notably liquefied petroleum gas (LPG). Starting around 2013, the Pratyaksha Hastaantarit Laabh (PAHAL) program revamped the government’s subsidy program for LPG. Instead of purchasing LPG in the market at a subsidized price, users received direct electronic payments linked to their bank accounts, preventing misuse of funds and allowing LPG distributors and program implementers to keep track of LPG connections through a centralized database. In 2015, the ‘Give it Up’ (GIU) campaign was launched, as 11 million LPG users voluntarily donated their LPG subsidies to give to poorer households. The GIU program leveraged the digital infrastructure created by the PAHAL scheme, allowing individual donors to look up on the GIU website which beneficiary received their subsidies. Building on the public momentum of the GIU campaign, in May 2016, the Ministry of Petroleum and Natural Gas (MoPNG) rolled out the Pradhan Faster Adoption and Manufacturing of (Hybrid &) Electric Vehicles in India (FAME India) Scheme was launched in 2015 to promote hybrid and electric 53 vehicles in the country. Phase II of the Scheme has been approved with an allocation of INR 10,000 crores to provide financial incentives for adoption and to establish a network of charging stations. 46 | CLEARING THE AIR: A TALE OF THREE CITIES Mantri Ujjwala Yojana initiative, with the goal of providing free LPG connections to 50 million poor households by 2019. Under the program, women were the only ones eligible to receive new connections, which were linked to their individual bank accounts and ID cards. The Ujjwala initiative reached its target ahead of schedule, so in 2018 the MoPNG decided to scale up its target even more to connect 80 million households by the end of 2020. As of November 2018, 57 million households had already received LPG connections. Evidence of the Ujjwala initiative’s success is plain to see. A large-scale field survey carried out in six states in central and northern India in 2018 by researchers with the Council on Energy, Environment, and Water (CEEW) showed that LPG use had expanded from 22 percent of households to 58 percent by 2018.54 About half of those households that had recently started using LPG credited the Ujjwala initiative as the reason for them getting a connection. The poorest households saw rates of LPG usage increase the fastest between 2015 to 2018 compared to the general population. Still, challenges remain with the use of LPG for cooking, including affordability, reliability, and convenience. According to the data from the CEEW survey, about 55 percent of households that received new LPG connections through the Ujjwala program continued to use traditional biomass as their primary fuel for cooking. The longer households use LPG, the more they are likely to rely on it for their primary fuel, so it could be that many of the Ujjwala participants are still in the phase of transitioning. Yet, LPG users who continued to cook with biomass commonly frequently said that buying fuel was not affordable, or that procuring fuel was difficult and time consuming. Other than in West Bengal, more than half of all households cannot receive LPG cylinders delivered to their homes. Most travel four kilometers or more to have their cylinders refilled, and most only have a single cylinder (while urban users typically have two). Of those who reportedly did not use LPG to cook, about 90 percent cited the expense of fuel as the main reason. Ultimately, by solving problems related to affordability, reliability, and convenience, the goal will be for households to move toward using gaseous fuels and electricity as their exclusive source of energy for cooking. Indeed, research has found that continuing to use traditional biomass even a small fraction of the time may still result in adverse health effects. In Ecuador, for example, which has provided large LPG price subsidies for decades, resulting in more than 90 percent of households using LPG as the primary cooking fuel, a recent study found that biomass continued to be used in rural areas, with two drivers being that biomass was frequently free and could also heat homes at the same time (Gould et al. 2018). Sources: Smith (2018); Jain et al. (2018); household air pollution health impacts data from Institute for Health Metrics and Evaluation (IHME), “Global Burden of Disease Study 2017 (GBD 2017) Results” (2018), http://ghdx.healthdata.org/gbd-results-tool; data on household access to clean fuels and cooking technologies from World Bank, “Sustainable Energy for all (SE4ALL)” database, available at https://databank.worldbank.org; email correspondence from Masami Kojima, 20 February 2019. 34. These efforts at the national and state level are seen in the decline of PM10 from 2016 to 2018 (Figure showing some limited impact on the ground. As 12c). Pollution levels, though, remain well above healthy shown in Figure B1.1, with air quality improvements in levels and much remains to be done to clear the air in some regions. Air quality in Delhi is also improving as Delhi and a growing number of Indian cities. 54 The states are Bihar, Jharkhand, Madhya Pradesh, Odisha, Uttar Pradesh, and West Bengal. CLEARING THE AIR: A TALE OF THREE CITIES | 47 What lessons can other countries draw? What lessons can other countries draw? 1. The experiences of these three cities suggest that FIGURE 14: Key Components of Air Quality Strategy there is no silver bullet. Improving air quality takes time, as solutions require changes to policies and to institutions, and sustained political commitment over many years is required. Substantial reductions in pollution are possible over a relatively short period with persistence, as Information are rebounds in pollution if enforcement is relaxed and commitment wanes. Mexico City began implementing policies to reduce air pollution in 1990. As a result, NO2, O3, Pb, and SO2 concentrations declined dramatically Air Quality in the 1990s. Mean annual concentrations of PM2.5, for Management which monitoring data are only available since in the mid- 2000s, also declined in the early 2000s, but rebounded in 2010s and remain above the national standard. Incentives Institutions Beijing similarly began efforts to tackle air pollution in 1990s, and finally saw some dramatic declines between 2013 and 2018. Air pollution remains a challenge and average levels in Beijing remained two times higher than the national standard and seven times higher than the WHO guideline values. Delhi saw a brief period of stable pollution concentration levels in early 2000s having air quality levels and raising awareness on the health and begun its air pollution management efforts in the early other economic costs of air pollution have been found to 1990s. Air quality however steadily declined in the 2000s, increase demand for action. In Mexico City, for example, building up to a crisis in the late 2010s. With concerted careful analysis of the impacts of air pollution on the efforts of the federal and the Delhi governments, health of children helped galvanized public support for Delhi appears to be turning the corner again on the city’s first air quality management strategy. In Delhi pollution levels. too, overall public support and awareness about the health impacts of air pollution contributed to the government 2. The experiences of these three cities also point to throwing its support behind the CNG conversion information, incentives, and institutions as the three program (Bell et al. 2004). And India’s National Air prongs of an effective air pollution management Quality Index program initiated in 2015 is an important strategy. As noted above, and interestingly, all three cities step towards supporting action. Additionally, apart began their programs of air quality management in the from better monitoring data, it is important to improve early 1990s but ended up in very different places in late data to inform air quality action plans, which, in turn, 2010s, pointing to differences in policies and programs require data on pollution sources and cost effectiveness to tackle air pollution. Three features of programs and of different policy interventions. Data on emissions and policies – adequacy and accessibility of information, sources of pollution allow policymakers to build models incentives for compliance, and institutions fit-for-purpose to assess expected improvements from current and – stand out and offer lessons to other cities. planned policy interventions, to set measurable targets and to identify strategies to meet targets in a cost-effective Information: adequate and accessible manner, and to monitor progress. Finally, timely and accessible data can support enforcement of regulations. 3. Data on air pollution concentrations and its health Installing emissions monitors in large industrial facilities implications, on sources of pollution, on enforcement, and power plants and making these data public can help etc. are critical to the design and implementation of hold local regulators and plant operators accountable for air quality programs. Expanding air quality monitoring upholding environmental standards, as has been the case networks, supporting public disclosure of information on in China. India’s policy requiring polluting industries CLEARING THE AIR: A TALE OF THREE CITIES | 49 to install Online Continuous Emissions/Effluent of emissions such that new sources are granted permits to Monitoring Systems will similarly improve enforcement establish and operate only if they agree to offset every unit of existing regulations. of emission from the unit by reduction of two units of emission elsewhere, a requirement that imposes a heavy Incentives: mainstreamed cost on new facilities and discourages development. US 4. Countries need a strong regulatory mechanism Clean Air Act also provides a carrot: section 105 of the to ensure that states and cities are incentivized to Act authorizes the federal government to provide grants implement policies and programs to reduce air equal up to 60 percent of the cost of state air quality pollution. Be it carrot- or stick-based, a mechanism to management programs. Currently federal funds on incentivize implementation of air quality management average provide 25 percent of the funding needs of state plans is needed. Cities other than Delhi -- Bengaluru, air programs. The recently announced performance- Hyderabad, Chennai, Ahmedabad, Kanpur, and Sholapur based grants to Indian cities for air quality improvements -- developed air quality management plans in 2003-04 to is a step in the right direction to create a mechanism reduce air pollution, at the behest of the Supreme Court. to incentivize cities to act. In China, local government The 11th Five-Year Plan (2007-2012) also provided for officials and state-owned enterprise managers were held cities to develop action plans to manage air pollution accountable for implementing air quality management (CSE 2016). Given the continued deterioration of air plan by integrating air quality targets and measures into quality, these plans did not lead to any meaningful official performance reviews. Officials that consistently reduction in air pollution, in part because of a lack of failed to meet targets were demoted and reassigned to a mechanism to incentivize compliance. Additionally, less prestigious jobs. Though effective, such a mechanism an examination of the role of the government and is not necessarily transferrable to other countries with Supreme Court in efforts to clean Delhi’s air points to different governance structures. a pattern of dependence on the courts for compliance. Time and time again, the government announced 5. While enforcement of regulations is essential, it measures to reduce pollution but did not follow is insufficient; incentives should also be provided to through on implementation. The Supreme Court then support compliance, and these can entail substantial weighed in to force the government to implement the fiscal outlays. In China, between 2013 and 2017, the policy measures it had previously announced. Should central government provided US$ 9.29 billion in special the Supreme Court continue to play this role, or can a funds and budgetary resources to support air quality stronger regulatory framework provide a mechanism to management in the region including Beijing and incentivize governments to implement policies designed the surrounding provinces and cities. These financial to tackle air pollution? Sanction powers granted to the resources were used to support a variety of incentive United States Environment Protection Agency (EPA) programs, including subsidies for end-of-pipe controls under the United States’ Clean Air Act (CAA) offer some and boiler retrofits in power plants and factories, rebates lessons to countries to incentivize implementation. As for scrapping older vehicles, and payments to households per the provisions of the CAA, if an area or city is found switching out coal-fired heating stoves for gas or electric to have pollution levels above acceptable standards, systems. Provinces, moreover, committed their own they are required to prepare and submit to the EPA a resources and used own and centrally allocated funds to State Implementation Plan (SIP). The SIP provides a leverage additional financing from the private sector in time bound set of measures, potentially to be imposed the order of US$ 2.96 billion. In the mid-2000s, Mexico on industry, transportation, etc., that are necessary to City provided direct subsidies to drivers of old taxis in achieve compliance with air quality standards. The CAA, exchange for retiring and scrapping their old vehicles, however, includes additional provisions to enforce SIP along with access to low-cost loans for vehicle renovations implementation. Namely, if the state fails to submit an or purchase of more efficient vehicles. Similarly, a range acceptable plan or fails to implement the measures of an of incentives were offered to encourage industrial approved plan, the CAA empowers EPA to impose one enterprises to make the switch from fuel oil to natural of two sanctions: (i) withholding certain federal highway gas and to install emissions control equipment. Fiscal funds by prohibiting the Secretary of Transportation from incentives and exemptions from emergency restrictions awarding funds from the Federal-aid Highway Program; that require industrial plants to curtail their production or (ii) imposing a “2:1 offset” requirement on new sources when air pollution reaches high level were included. The 50 | CLEARING THE AIR: A TALE OF THREE CITIES Government of India provided US$ 160 million between 7. Air pollution management strategies need to be 2018 and 2019 to subsidize the distribution of machinery integrated into multi-sector development plans, for in-situ management of crop residue to farmers in the and an institutional set up is similarly required to states of Punjab, Haryana, and Uttar Pradesh. facilitate this, to match the cross-sectoral nature of the air pollution challenge. Addressing air pollution Institutions: fit-for-purpose requires both national-level policies and local action. Examples of national policies include regulations, 6. The multi-jurisdictional nature of air pollution standards, sectoral policies, and supporting measures requires an institutional setup that reaches across that incentivize implementation, such as taxes or individual jurisdictions – an airshed-based55 subsidies. Local actions include planning, monitoring, management approach. Because air pollution travels enforcement, and investments. Moreover, to be effective, across administrative boundaries, and pollution sources air quality management activities need to be embedded are located both inside and outside any given city, an in national and state development plans, and not just in airshed-based management approach that cuts across standalone air quality management plans. Notably, three jurisdictions is essential to achieving results. In other significant sources of pollution in India – residential words, to effectively address the sources of pollution, biomass burning, agricultural emissions, and dust – do air quality should be managed at the same scale as the not fall under the direct purview of pollution boards. problem. The JJJ Regional Air Quality Prevention and Program such as the Pradhan Mantri Ujjwala Yojana (see Control Coordination Group was established in China Box 4) that has expanded access to clean cooking fuels, to achieve cross-jurisdictional coordination. The group most notably LPG, reducing the reliance on residential has high-level participation from all administrative biomass burning, is essential to efforts to reduce air entities in the JJJ region, including the Beijing City pollution. Such a program goes well beyond the mandate Governor, Tianjin City Governor, and Hebei Provincial of pollution control board and gets to the heart of how Governor, as well as leading officials from the relevant development programs are designed. Similarly, reducing sectoral ministries, including the Ministry of Housing emissions from power generation and small and medium and Urban Development, Ministry of Transportation, enterprises will entail increasing the use of natural gas and Ministry of Agriculture, and so on. The group is led by the renewable energy and goes well beyond the provisions State Council, China’s highest governmental body. The of the Indian Air Act to prescribe and enforce emission group is responsible for formulating targets and annual standards for power plants and industry. In China, for implementation plans for air quality management across example, the Ministries of Environmental Protection, the JJJ region and to set policies for cross-jurisdictional Industry and Information Technology, Finance, Housing issues such as fuel standards, energy supply, and public and Rural Development, along with the National transportation. Provincial and city-level governments Development and Reform Commission and National continue to be the primary implementers for these air Energy Administration, joined together to issue a five- quality management programs, however. A similar role is year action plan for air pollution prevention and control played by the Megalopolis Environmental Commission for the entire JJJ airshed. Their joint efforts led to the in Mexico, which brings together federal authorities from dramatic reduction in coal use in the JJJ region. the ministries of environment, health, and transport with local authorities from Mexico City and 224 municipalities 8. In summary, air pollution is a critical challenge from the neighboring states of Mexico, Hidalgo, for many countries. It is today a major health risk Morelos, Puebla, and Tlaxcala, which jointly define an and a drag on their economies. But it is a challenge airshed for Mexico City. Lessons from these and other that can nevertheless be tackled with sustained political countries with effective intergovernmental coordination commitment, policy innovations, and fit-for-purpose suggest that such mechanisms are successful if they are: institutional arrangements. At the same time, it is a (i) executive driven; (ii) housed within central agencies at challenge that is best addressed now, before countries put the federal level or an autonomous institution; and that in place their built environment and make energy mix are (iii) formalized to ensure that governments interact choices, among other development choices, that will lock with one another on a regular basis. them into a more or less pollution intensive growth path. 55 An airshed is a part of the atmosphere that behaves in a coherent way with respect to the dispersion of emissions. CLEARING THE AIR: A TALE OF THREE CITIES | 51 References and Annexure References Amman, M., P. Purohit, I. Bertok, J. Borken-Kleefeld, J. Cofala, C. Heyes, G. Kiesewetter, Z. Klimont, L. Jun, B. Ngyuen, P. Rafaj, R. Sander, W. Schopp, A.D. Bhanarkar, P. S. Rao, A. Shrivastava, D. Majumdar, B. Harsh Vardhan. 2016. Managing future air quality in Delhi. National Environmental Engineering Research Institute, Nagpur, India, and International Institute for Applied Systems Analysis, Austria. ARAI-TERI (The Automotive Research Association of India and The Energy and Resources Institute). 2018. “Source Apportionment of PM2.5 and PM10 of Delhi NCR for Identification of Major Sources.” Report to Department of Heavy Industry, Ministry of Heavy Industries and Public Enterprises, New Delhi, India, August 2018. Balikrishnan, Kalpana et al. (India State-Level Disease Burden Initiative Air Pollution Collaborators). 2018. “The Impact of Air Pollution on Deaths, Disease Burden, and Life Expectancy Across the States of India: The Global Burden of Disease Study 2017.” The Lancet Planetary Health (December 6), doi: 10.1016/S2542-5196(18)30261-4. Bell, R. G., K. Mathur, U. Narain, and D. Simpson. 2004. Clearing the Air: How Delhi Broke the Logjam on Air Quality Reforms. Environment Science and Policy for Sustainable Development 46(3):22-39. Blackman, A., F. Alpízar, F. Carlsson and M. Rivera. 2018. “A Contingent Valuation Approach to Estimating Regulatory Costs: Mexico’s Day Without Driving Program.” Journal of the Association of Environmental and Resource Economists 5(3): 607–641. Burney, Jennifer and V. Ramanathan. 2015. “Recent Climate and Air Pollution Impacts on Indian Agriculture.” Proceedings of the National Academy of Sciences 111, no. 46: 16319-324. CEA (Central Electricity Authority, Ministry of Power, Government of India). 2018. “Quarterly Review Report, Renovation and Modernization of Thermal Power Stations.” CEA, New Delhi, September 2018. CEEW (Council on Energy, Environment and Water) - IIASA (International Institute for Applied Systems Analysis). 2018. Pathways to Achieve National Ambient Air Quality Standards in India. March 2019. Chakraborty, Abhishek and Tarun Gupta. 2010. “Chemical Characterization and Source Apportionment of Submicron (PM1) Aerosol in Kanpur Region, India.” Aerosol and Air Quality Research 10: 433–445. Chang, Tom, Joshua Graff Zivin, Tal Gross, and Matthew Neidell. 2016. “The Effect of Pollution on Worker Productivity: Evidence from Call-Center Workers in China.” National Bureau of Economic Research Working Paper 22328, Cambridge, MA, United States, June 2016. Chen, Fei, Hongxia Yu, Jun Hu, and Fahe Chai. 2017. “Analysis of Sources of PM2.5 in Xingtai Using Chemical Mass Balance Model” (in Chinese). Journal of Ecology and Rural Environment 33, no. 12: 1075-1083. CSE (Centre for Science and Environment). 2016. Legal Framework for Clean Air in Cities: Gaps and Potential. Centre for Science and Environment, New Delhi. Cusworth, Daniel, Loretta Mickley, Melissa Sulprizio, Tianjin Liu, Miriam Marlier, Ruth DeFries, Sarath Guttikunda, and Pawan Gupta. 2018. “Quantifying the Influence of Agricultural Fires in Northwest India on Urban Air Pollution in Delhi, India.” Environmental Research Letters 13, DOI: 10.1088/1748-9326/aab303. Dasgupta, Susmita, Benoit Lapante, Hua Wang, and David Wheeler. 2002. “Confronting the Environmental Kuznets Curve.” Journal of Environmental Perspectives 16, no. 1: 147-168. . Journal of Political Economy 116, 1: 38–79. Davis, L. W. 2008. “The Effect of Driving Restrictions on Air Quality in Mexico City” Eaton, Sarah and Genia Kostka. 2014. “Authoritarian Environmentalism Undermined? Local Leaders’ Time Horizons and Environmental Policy Implementation in China.” The China Quarterly 218: 359-380. EPA (U.S. Environmental Protection Agency). 2009. “Integrated Science Assessment for Particulate Matter.” http://cfpub.epa.gov/ ncea/risk/recordisplay.cfm?deid=216546. EPCA (Environment Pollution (Prevention and Control) Authority). 2019. Note on Constitution and Functioning of EPCA by Amicus Curiae dated 26.8.2019. http://www.epca.org.in/EPCA-Reports1999-1917/Note-on-the-Constitution-and-Functioning-of- EPCA.pdf. EPCA-CSE. 2018. Air Pollution Report Card 2017-18. A status report by Environment Pollution (Prevention & Control) Authority for Delhi and Centre for Science and Environment. February 28, 2018. CLEARING THE AIR: A TALE OF THREE CITIES | 53 Eskeland, G. S., and T. Feyzioglu. 1995. “Rationing Can Backfire. The Day without a Car in Mexico City.” Policy Research Working Paper 1554. Washington, D.C.: The World Bank. EU (European Union). 2017. Commission Implementing Decision (EU) 2017/1442 of 31 July 2017. Establishing Best Available Techniques (BAT) Conclusions, Under Directive 2010/75/EU of the European Parliament and of the Council, for Large Combustion Plants. Official Journal of the European Union, 17 August 2017, L 212. GBD 2016 Risk Factors Collaborators. 2017. “Global, Regional, and National Comparative Risk Assessment of 84 Behavioural, Environmental and Occupational, and Metabolic Risks or Clusters of Risks, 1990-2016: A Systematic Analysis for the Global Burden of Disease Study 2016.” The Lancet 2017 390: 1345- 1422. GoI (Government of India). 2018. National Clean Air Programme (NCAP)-India. GoM (Government of Mexico). 1990. Metropolitan Commission for the Prevention and Control of Pollution in the Valley of Mexico. Comprehensive Program against Air Pollution: A Common Commitment. Department of the Federal District (in Spanish). http:// www.aire.cdmx.gob.mx/descargas/publicaciones/gestion-ambiental-aire-memoria-documental-2001-2006/descargas/programa_ integral_contra_la_contaminacion_atmosferica.pdf. GoM (Government of Mexico). 1996. Program to Improve Air Quality in the Valley of Mexico 1995-2000 (in Spanish). GoM (Government of Mexico). N.d. Program to Improve Air Quality in the Metropolitan Zone of the Valley of Mexico 2011-2020 (in Spanish). http://www.aire.cdmx.gob.mx/descargas/publicaciones/flippingbook/proaire-2011-2020-anexos/ Gould, Carlos F., Samuel Schlesinger, Andres Ochoa Toasa, Mark Thurber, William F. Waters, Jay P. Graham, and Darby W. Jack. 2018. “Government policy, clean fuel access, and persistent fuel stacking in Ecuador.” Energy for Sustainable Development 46: 111–122, https://doi.org/10.1016/j.esd.2018.05.009. Grossman, Gene and Alan Krueger. 1991. “Environmental Impacts of a North American Free Trade Agreement.” Working Paper No. 3914, National Bureau of Economic Research (NBER), Cambridge, United States, November 1991. IARC (International Agency for Research on Cancer). 2013. “Air Pollution and Cancer.” IARC Scientific Publication No. 161. http:// www.iarc.fr/en/publications/books/sp161/. ICCS (Innovation Center for Clean-Air Solutions). 2018. “Assessment of Jing-Jin-Ji Air Quality Improvement (2013-2017). Draft report to World Bank, Beijing, China, June 2018. IEA. 2018a. World Energy Outlook 2018. Paris, OECD. IEA. 2018b. World Energy Statistics and Balances. Paris, OECD. IHME (Institute for Health Metrics and Evaluation, University of Washington, Seattle). 2017. “Global Burden of Disease Study 2016 (GBD 2016) Results.” Global Burden of Disease Collaborative Network, IHME, http://ghdx.healthdata.org/gbd-results-tool. INSP-UNICEF Mexico. 2016. National Survey of Boys, Girls, and Women 2015 (Survey of Multiple Indicators by Cluster 2015) – Final Report. Mexico City, Mexico: National Institute of Public Health and UNICEF Mexico. Jain, Abhishek, Saurabh Tripathi, Sunil Mani, Sasmita Patnaik, Tauseef Shahidi, and Karthik Ganesan. “Access to Clean Cooking Energy and Electricity: Survey of States 2018.” Report by Council on Energy, Environment, and Water (CEEW), November 2018. James, Christopher. 2017. “China’s Power Sector and Air Quality Reforms: Global Lessons on Getting Institutional Responsibilities Right.” Regulatory Assistance Program, Beijing, China, 7 November 2017, https://www.raponline.org/knowledge-center/chinas- power-sector-air-quality-reforms-global-lessons-getting-institutional-responsibilities-right/?sf_action=get_data&sf_data=results&_ sft_region=china&_sft_language=english. Kostka, Genia. 2015. “Command Without Control: The Case of China’s Environmental Target System.” Regulation and Governance, DOI:10.1111/rego.12082. Kummu, Matti, Maija Taka, and Joseph Guillaume. 2018. “Gridded Global Datasets for Gross Domestic Product and Human Development Index Over 1990-2015.” Nature: Scientific Data 5, DOI: 10.1038/sdata.2018.4. Kuznets, Simon. 1955. “Economic Growth and Income Inequality.” The American Economic Review 45, no. 1: 1-28. Lei, W., G. Li, and L.T. Molina. 2013. “Modeling the Impacts of Biomass Burning on Air Quality in and around Mexico City.” Atmospheric Chemistry and Physics 13: 2299-2319. McKinley, Galen, Miriam Zuk, Morton Höjer, Montserrat Avalos, Isabel González, Rodolfo Iniestra, Israel Laguna, Miguel Martínez, Patricia Osnaya, Luz Reynales, Raydel Valdés, and Julia Martínez. 2005. “Quantification of Local and Global Benefits from Air Pollution Control in Mexico City.” Environmental Science and Technology 39, no. 7: 1954-1961. 54 | CLEARING THE AIR: A TALE OF THREE CITIES MENR (Ministry of Environment and Natural Resources) 2017. Federal Management Program to Improve the Air Quality of the Megalopolis: PROAIRE of the Megalopolis 2017-2030 (in Spanish). MEP (Ministry of Environmental Protection, Government of China). 2012. Notice Regarding Implementation of Ambient Air Quality Standards (GB 3095-2012) (in Chinese), 29 February 2012. MESM-MEFD-MENR-MH (Ministry of Environment of the Government of the State of Mexico, Ministry of the Environment of the Federal District, Ministry of the Environment and Natural Resources and Ministry of Health). 2011. Programa para mejorar la calidad del aire de la Zona Metropolitana del Valle de México 2011-2020. http://dsiappsdev.semarnat.gob.mx/datos/portal/proaire/11_ ProAire%20ZMVM.pdf Mexico City. 1990. Integrated Program to Control Atmospheric Pollution – A Shared Promise (in Spanish), October 1990. MoEFCC (Ministry of Environment, ForestBe, and Climate Change). 2018. EPCA Reconstituted order dated 03rd October 2018. http://epca.org.in/EPCA-reconstitution-order-dt.3Oct2018.pdf Molina, Luisa and Mario Molina. 2006. “Improving Air Quality in Megacities: Mexico City Case Study.” Annals of the New York Academy of Sciences, DOI: 10.1196/annals.1319.006. NIECC (National Institute of Ecology and Climate Change). 2019. Environmental Impact of the Sulfur Content in Commercial Diesel Vehicles in Mexico (in Spanish). General Coordination of Pollution and Environmental Health. https://www.gob.mx/cms/ uploads/attachment/file/457768/Azufre_en_diesel_y_emisiones.pdf. NISGI (National Institute of Statistics, Geography and Informatics). 2005. Statistics of the Environment of the Federal District and Metropolitan Area 2002 (in Spanish). http://www.paot.org.mx/centro/inegi/amdf2002/archivo10.pdf. Narain, Urvashi and Christopher Sall. 2016.“Methodology for Valuing the Health Impacts of Air Pollution: Discussion of Challenges and Proposed Solutions.” World Bank, Washington, DC, 16 June 2016. Payanotou, Theodore. 1997. “Demystifying the Environmental Kuznets Curve: Turning a Black Box into a Policy Tool.” Environment and Development Economics 2: 465-484. PAMC (Public Administration of Mexico City) 2019. Notice on Program to Prevent and Respond to Atmospheric Environmental Contringencies in Mexico City. Official Gazatte of Mexico City (in Spanish). May 28, 2019. http://www.aire.cdmx.gob.mx/descargas/ ultima-hora/calidad-aire/pcaa/Gaceta_Oficial_CDMX.pdf SESM (Secretary of Environment of State of Mexico). 2019. Agreement establishing the Program for Attention to Ambient Contingencies in the Metropolitan Area of the Valley of Toluca and the Metropolitan Area of Santiago Tianguistenco. Government Gazatte (in Spanish). May 28, 2019. http://legislacion.edomex.gob.mx/sites/legislacion.edomex.gob.mx/files/files/pdf/gct/2019/ may282.pdf SEMARNAT. 2010. Agreement establishing the Program for the Reduction of Pollutant Emissions to the Atmosphere. Official Journal of the Federation (in Spanish). May 19, 2010. https://dof.gob.mx/nota_detalle.php?codigo=5143329&fecha=19/05/2010 Shaddick, G., M.L. Thomas, A. Jobling, M. Brauer, A. van Donkelaar, R. Burnett, H.H. Chang, A. Cohen, R. Van Dingenen, C. Dora, S. Gumy, Y. Liu, R.V. Martin, L.A. Waller, J. West, J.V. Zidek, and A. Prüss-Ustün. 2018. "Data Integration Model for Air Quality: A Hierarchical Approach to the Global Estimation of Exposures to Ambient Air Pollution." Applied Statistics Series C 67, no. 1: 231- 253. Shafik, Nemat and Sushenjit Bandyopadhyay. 1992. “Economic Growth and Environmental Equality.” World Bank Policy Research Working Paper 904, Background Paper for World Development Report 1992, World Bank, Washington, DC, June 1992. Sharma, M. and O. Dikshit. 2016. Comprehensive Study of Air Pollution and Green House Gases (GHGs) in Delhi (Final Report: Air Pollution component) accessed at http://delhi.gov.in/DoIT/Environment/PDFs/Final_Report.pdf on 17th February 2017. Shi, Guoliang, Jiayuan Liu, Haiting Wang, Yingze Tian, Jie Wen, Xurong Shi, Yinchang Feng, Cesunica E. Ivey, and Armistead G. Russell. 2018. “Source Apportionment for Fine Particulate Matter in a Chinese City Using an Improved Gas-Constrained Method and Comparison with Multiple Receptor Models.” Environmental Pollution 233: 1058-67 Smith, Kirk. 2018. “Pradhan Mantri Ujjwala Yojana: Transformation of Subsidy to Social Investment in India” in Making of New India: Transformation under Modi Government, ed. Bibek Debroy, Anirban Ganguly, and Kishore Desai, 401-411. New Delhi: Wisdom Tree. Stanaway, Jeffrey et al. (GBD 2017 Risk Factor Collaborators). 2018. “Global, Regional, and National Comparative Risk Assessment of 84 Behavioral, Environmental and Occupational, and Metabolic Risks or Clusters of Risks for 195 Countries and Territories, 1990-2017: A Systematic Analysis for the Global Burden of Disease Study 2017.” The Lancet 392: 1923-94. CLEARING THE AIR: A TALE OF THREE CITIES | 55 Stern, D. I. 2015. The environmental Kuznets curve after 25 years. Crawford School of Public Policy, Australian National University, Australia. UNEP (United Nations Environment Programme). 2016. “A Review of Air Pollution Control in Beijing: 1998-2013.” UNEP, Nairobi, Kenya, May 2016. Unruh, Gregory and William Moomaw. 1998. “An Alternative Analysis of Apparent EKC-Type Transitions.” Ecological Economics 25: 221-229. van Donkelaar, A., R.V. Martin, M. Brauer, N.C. Hsu, R.A. Kahn, R.C. Levy, A. Lyapustin, A.M. Sayer, and D.M. Winker. 2016. "Global Estimates of Fine Particulate Matter using a Combined Geophysical-Statistical Method with Information from Satellites, Models, and Monitors." Environmental Science and Technology 50, no. 7: 3762–3772. Vega, Elizabeth, Silvia Eidels, Hugo Ruiz, Diego López-Veneroni, Gustavo Sosa, Eugenio Gonzalez, Jorge Gasca, Virginia Mora, Elizabeth Reyes, Gabriela Sánchez-Reyna, Rafael Villaseñor, Judith Chow, John Watson, and Silvia Egerton. 2010. “Particulate Air Pollution in Mexico City: A Detailed View.” Aerosol and Air Quality Research 10: 193-211. Wang, Alex. 2013. “The Search for Sustainable Legitimacy: Environmental Law and Bureaucracy in China.” Harvard Environmental Law Review 37, no. 2: 365-440. WHO (World Health Organization). 2005. “WHO Air Quality Guidelines for Particulate Matter, Ozone, Nitrogen Dioxide, and Sulfur Dioxide – Global Update 2005: Summary of Risk Assessment.” Document WHO/SDE/PHE/OEH/06.02, WHO, Geneva, Switzerland. WHO. 2013. “Review of Evidence on Health Aspects of Air Pollution - REVIHAAP Project: Final Technical Report.” December 2. http://www.euro.who.int/en/health-topics/environment-and-health/air-quality/publications/2013/review-of-evidence-on-health- aspects-of-air-pollution-revihaap-project-final-technical-report. WHO. 2018. “WHO Global Ambient Air Quality Database (Update 2018),” http://www.who.int/airpollution/data/cities/en/. WHO-UNEP. 1992. Urban Air Pollution in the Mega Cities of the World. World Health Organization, United Nations Environment Programme, Blackwell, Oxford. World Bank. 2016. “Program Appraisal Document on a Proposed Loan in the Amount of US$ 500 Million to the People’s Republic of China for a Hebei Air Pollution Prevention and Control Program.” Washington, DC, United States, May 2016, http://projects. worldbank.org/P154672?lang=en. World Bank-IHME (World Bank and Institute for Health Metrics and Evaluation, University of Washington, Seattle). 2016. The Cost of Air Pollution: Strengthening the Economic Case for Action. World Bank, Washington, DC, 8 September 2016. Zhang, Xing. 2016. “Emissions Standards and Control of PM2.5 from Coal-Fired Power Plant.” IEA Clean Coal Centre Report CCC/267, London, July 2016. Zhu, F. and L. Wang. 2014. “Analysis on Technology-Economy and Environment Benefit of Ultra-Low Emission from Coal-Fired Power Units.” Environmental Protection 41, no. 21: 28-33. Annex Pollution Intensity of Economic Growth in South Asia and Other Regions Econometric Analysis Reveals Economic Growth in to account for possible differences in energy needs South Asia Systematically More Pollution-Intensive for heating and cooling, which may result in greater Than in Other Regions emissions and poorer air quality, particularly in areas where households do not have access to clean fuels and technologies for heating and cooking. Countries where Econometric analysis was performed to test the large areas are classified as arid (that is, with less than hypothesis that economic growth has been systematically 250 mm precipitation annually) may experience more more pollution-intensive in India and the other South problems with windblown dust. The length of coastline Asian countries than it has for countries at comparable and mean TRI reflect differences in topographic and levels. Adapting an approach developed by Stern et al. geographic characteristics that might influence the (2016), the relationship between the long-run growth in diffusion of PM2.5 pollution. pollution and income per capita as: Time-varying controls account for changes in the = 0 + 1 + 1 ,0 + 2 ,0 + 3 ,0 + =4 + (1) variables, , as well as the initial levels, X_(i,0). Time- varying controls include: (a) the share of coal, oil, where the subscript indexes countries. The dependent and primary solid biofuels in total primary energy variable is the average annual rate of change in consumption; (b) population density; (c) the share of pollution,56 calculated for the period from year 0 to year gross value added (GVA) by manufacturing in total GDP; T as = ( , ,0 ) / ; is the average annual growth (d) the share of GVA by agriculture, forestry, and fishing rate of GDP per capita (year 2011 US$, market rates), in total GDP; and (e) the degree of outward integration calculated similarly to ; ,0 is the initial level of GDP per and openness of a country’s economy, measured as the capita in year 0; ,0 is the initial level of pollution in year ratio of exports to GDP. (a) will result in greater emissions 0; Xi is a vector of control variables; and εi is an error term. intensity of energy and therefore worse air quality. Higher population density is also likely to result in worse air The vector of controls, Xi, includes both time-invariant quality, given a higher household demand for energy and and time-varying variables. Time-invariant variables a greater concentration of pollutant-emitting economic include: (a) the minimum temperature of the coldest activities. Controls (c) and (d) account for differences month; (b) the maximum temperature of the warmest in the structure of economic output. Manufacturing is month; (c) the length of coastline as a ratio of the assumed to be the most emissions-intensive sector of the length of a country’s total perimeter; (d) the mean economy, so it is expected that (c) will translate into more terrain ruggedness index (TRI) score for urban areas in pollution. Agriculture (d) may also contribute to PM2.5 the country; and (e) the share of a country’s land area emissions, including through burning crop stubble, and with mean annual precipitation of less than 250 mm. contributions to secondary particle. We are agnostic Minimum and maximum temperatures are included In the Stern et al. (2016) study, p is the log of emissions per capita. In our analysis, we instead use the log of population-weighted mean annual PM2.5 56 exposure. CLEARING THE AIR: A TALE OF THREE CITIES | 57 about the effect of (e), though researchers have included = + + + + + 0 1 1 ,0 2 ,0 3 ,0 this variable in previous studies to try to capture the potential effect of polluting, export-oriented industries 4 + 5 + 6 ,0 + =7 + (2) migrating to countries with lax environmental standards. The regional dummies are interacted with GDP per capita The timeframe of the analysis is 1995 to 2015, so and GDP growth, so the overall effect of GDP per capita T = 20. As in Stern et al. (2017), we subtract the across- growth on air quality is 1 + 1 ,0 + 5 for countries in sample means of all continuous levels variables before South Asia and 1 + 1 ,0 elsewhere.58 The coefficient estimating the equation. Thus, the average annual rate of of interest is β5. If β5 ≠ 0, then the income elasticity is change in PM2.5 for the “average” country in the sample is systematically different for countries in South Asia.59 given by 0 + 1 , where α1 is similar to a time effect and α1 is the income elasticity. The implied EKC turning Changes in air quality in one country may affect in downwind countries, though much of the earlier EKC point where pollution peaks occurs where = 0 as GDP literature ignored the spatial nature of pollution.60 To 1+ 1 0 capture the transboundary nature of air pollution, we per capita reaches τ = exp ( ) with y̅ 0 1 estimate equation (2) by introducing a spatial lag of the denoting the across-sample mean of ln (GDP per capita) dependent variable and allowing for spatial dependence in 1995. If τ occurs within the observed range of GDP per in the error terms. Rewriting the equation in column- capita for countries in the sample, this would support the vector notation: EKC hypothesis and the existence of a turning point in pollution levels as incomes rise. τ below the observed = 0 + 1 + 1 0 + 2 0 + + + range of GDP per capita would suggest that the pollution = + (3) declines inversely with income for all observed levels of income and τ above the observed range of GDP per where W is a matrix of spatial weights, proportional to capita would suggest that pollution is expected to worsen the inverse distance between countries; λW represents for all observed levels of income.57 the spillover effect from the estimated change in PM2.5 in neighboring countries; and ε is a spatially autoregressive To test if the income elasticity of pollution is systematically error term, with ρW accounting for the spillover of different in South Asian countries after controlling for the errors.61 other characteristics, we introduce a regional dummy variable r_i, equal to 1 if the country is in the South Asian region and 0 otherwise: 57 In the usual setup, the EKC hypothesis suggests that α1>0 and β1<0, meaning that pollution initially worsens with income growth and then eventually improves, as the elasticity of pollution with respect to income declines with the level of income. Because we have subtracted the sample mean from yi,0, however, this is not necessarily the case. Depending on the relative magnitude of α1 and β1, an inverted U-shape can also occur α1<0 and β1<0, as yi,0<0 for all yi,0 below the sample mean. What is important for the EKC hypothesis to hold is that β1<0. 58 We could also interact yi̇ yi,0ri, allowing for the rate at which the income elasticity of pollution declines with higher income to vary across regions. We omit this term, however, to make the interpretation of the coefficients and hypothesis testing more straightforward. 59 We also estimate equation (2) with a full set of dummies for the other regions and test if β5 is equal to the coefficients for the other regions. 60 This would include all the studies reviewed by Dasgupta et al. (2002) and Stern (2004). Examples of more recent studies that have taken a spatially-explicit approach include studies by Keene and Deller (2013) of ambient PM2.5 US counties and by Wang et al. (2018) of CO2 emissions in different regions of China. 61 Distances are measured between the centroids of countries. The W matrix is scaled so the largest eigen value equals 1. In theory, we would want data for all countries in the world to construct the W matrix. In practical terms, however, this is not very realistic. We estimate the model with data for 176 countries. Any countries with missing data are assumed to have zero influence on neighboring countries. Most of the countries with missing data are small island nations, for which it seems reasonable to assume zero influence on downwind countries; however, there are a few large countries with missing data. These include Eritrea, Finland, Niger, and South Sudan. We do have data for all the South Asian countries. 58 | CLEARING THE AIR: A TALE OF THREE CITIES Equation (3) is estimated using generalized spatial two- Two main results are as follows: stage least squares (GS2SLS) and maximum likelihood (ML) methods.62 The regression results are provided in First, the signs of the estimated EKC coefficients α̂ 1 and tables A.1 and A.2 below. Large differences in the results β̂ 1in the GS2SLS and ML models are consistent with the of the GS2SLS and ML estimates may suggest that the EKC hypothesis, although α̂ 1 is statistically insignificant. data are not independent and identically distributed. The implied EKC turning point at which mean annual Considering this, we estimate the GS2LS model for PM2.5 is expected to peak occurs with GDP per capita equation (3) with heteroskedastic-robust errors. We also somewhere between US$ 3,379 and US$ 4,578 for experiment with two different spatial weighting matrices. countries outside South Asia in the different models.63 In the first set of GS2SLS and ML regressions (“Model 1” in tables A.1 and A.2 ), we impose no restrictions on the Second, in all the models, the estimated coefficient maximum distance over which countries can influence of interest β̂ 5 is positive and statistically significant, each other. In the second set of regressions (“Model 2” meaning that the income elasticity of ambient PM2.5 is in tables A.1 and A.2), we set the values in the W matrix systematically higher in the South Asia region than for equal to zero if countries’ centroids are farther than 1,500 other regions. The implied EKC turning point for South km apart. Because of the spatial lag of the dependent Asia would occur at GDP per capita far above the sample variable , the coefficients in tables A.1 and A.2 are range—effectively meaning that there is no EKC turning difficult to interpret on their own and should not be point for South Asian countries if they continue along misconstrued as the direct effect of the variables on the their current growth path. outcome after eliminating spillover effects. 62 Large differences in the results of the GS2SLS and ML estimates may suggest that the data are not independent and identically distributed. Considering this, we estimate the GS2LS model for equation (3) with heteroskedastic-robust errors. We experiment with two different spatial weighting matrices. In the first case, we impose no restrictions on the maximum distance over which countries can influence each other. In the second case, we set the values in the W matrix equal to zero if countries are farther than 1,500 km apart. We also estimate equation (2) first using ordinary least squares (OLS) and apply the Moran’s I test to determine if spatial clustering exists in the residuals. The results of Moran’s I test confirm that spatial dependencies do exist. 63 Estimates of the EKC turning point here are for the second set of models in which spatial weights are restricted to countries with centroids within 1,500 km of each other. GDP per capita is taken in constant year 2010 US$ at market rates. Although GDP per capita measured at purchasing power parity (PPP) would be preferable, data are missing for too many countries to estimate the model for 1995 to 2015. CLEARING THE AIR: A TALE OF THREE CITIES | 59 TABLE A.1: Generalized spatial two-stage least squares (GS2SLS) regression results Dependent Variable: PM2.5 GS2SLS (se) GS2SLS (se) exposure growth rate (ṗi) Model 1 Model 2 GDP/person growth rate (ẏi) 0.019 (0.029) 0.001 (0.030) ln of GDP/person in 1995 (yi,0) 0.000 (0.001) 0.000 (0.001) ẏi yi,0 -0.016 (0.011) -0.023** (0.011) ln of PM2.5 in 1995 (pi,0 ) 0.000 (0.001) 0.000 (0.001) Coastline length / country perimeter (km/km) -0.001 (0.002) 0.000 (0.002) ln of urban terrain ruggedness index (TRI) 0.002* (0.001) 0.002** (0.001) Max temp of warmest month 0.001 (0.001) 0.002* (0.001) Min temp of coldest month -0.001 (0.001) -0.003** (0.001) Fraction of territory that is arid 0.006 (0.004) 0.007* (0.004) Change in fraction of energy from oil, coal, biofuels 0.005 (0.053) 0.026 (0.054) Fraction of energy from oil, coal, biofuels in 1995 0.001 (0.003) 0.001 (0.003) Population density growth rate 0.080 (0.075) 0.059 (0.078) ln of population density in 1995 -0.000 (0.001) 0.000 (0.001) Change in ratio of exports to GDP -0.019 (0.047) 0.007 (0.042) Ratio of exports to GDP in 1995 -0.002 (0.002) -0.000 (0.002) Change in share of manufacturing GVA in GDP 0.014** (0.007) 0.017** (0.007) Share of manufacturing GVA in GDP in 1995 (fraction) 0.007 (0.006) 0.011* (0.006) Change in share of agriculture GVA in GDP 0.008 (0.009) 0.008 (0.009) Share of agriculture GVA in GDP in 1995 (fraction) 0.007 (0.008) 0.010 (0.008) Country is in South Asia region (r_i) -0.034*** (0.011) -0.028*** (0.010) ẏi ri 0.385** (0.162) 0.326** (0.154) yi,0 ri -0.014*** (0.005) -0.013*** (0.005) Constant (α0) -0.001 (0.001) -0.000 (0.002) Coefficient on spatial lag of dependent variable (λ) 2.306*** (0.452) 0.678*** (0.253) Coefficient on spatially autoregressive error (ρ) 2.479*** (0.470) 2.536*** (0.326) Observations (countries) 176 176 Pseudo R2 -0.236 0.511 Notes and sources: (se) = heteroskedastic-robust standard errors; * significant at 90% level, ** significant at 95% level, *** significant at 99% level. Data on coastline length is from the CIA World Factbook. Data on maximum temperature, minimum temperature, and precipitation are from the WorldClim V1 Bioclim dataset (Hijmans et al.), obtained and processed in Google Earth Engine. Temperature variables have been standardized to have a mean of zero and standard deviation of one. Areas with less than 250 mm of annual precipitation are classified as arid. TRI is calculated following the method of Riley et al. (1999) using SRTM digital elevation model (v4) data, also obtained and processed through Google Earth Engine, and extents of built-up urban areas from Schneider et al. (2009), obtained from Natural Earth. Population data are from the UN’s World Population Prospects 2017 dataset. Share of agriculture and manufacturing gross value added (GVA) in GDP are calculated using official country national accounts data from UNdata.org and the UN Statistical Divisions’ National Accounts Main Aggregates database. Data on exports are from the World Bank World Development Indicators database. 60 | CLEARING THE AIR: A TALE OF THREE CITIES TABLE A.2: Maximum likelihood (ML) estimator regression results Dependent Variable: PM2.5 ML (se) ML (se) exposure growth rate (ṗi) Model 1 Model 2 GDP/person growth rate (ẏi) 0.024 (0.036) 0.008 (0.035) ln of GDP/person in 1995 (yi,0) 0.000 (0.001) 0.000 (0.001) ẏi yi,0 -0.021 (0.017) -0.023 (0.016) ln of PM2.5 in 1995 (pi,0) -0.001 (0.001) -0.000 (0.001) Coastline length / country area (km/km2) -0.001 (0.002) -0.001 (0.002) ln of urban terrain ruggedness index (TRI) 0.002** (0.001) 0.002** (0.001) Max temp of warmest month 0.002** (0.001) 0.002* (0.001) Min temp of coldest month -0.002** (0.001) -0.003*** (0.001) Fraction of territory that is arid 0.006* (0.003) 0.007** (0.003) Change in the fraction of energy from oil, coal, biofuels -0.031 (0.067) 0.007 (0.065) Fraction of energy from oil, coal, biofuels in 1995 0.002 (0.003) 0.002 (0.003) Population density growth rate 0.129** (0.060) 0.105* (0.062) ln of population density in 1995 0.000 (0.001) 0.000 (0.000) Change in ratio of exports to GDP 0.011 (0.059) 0.035 (0.057) Ratio of exports to GDP in 1995 -0.002 (0.003) -0.000 (0.003) Change in share of manufacturing GVA in GDP 0.013 (0.009) 0.015* (0.008) Share of manufacturing GVA in GDP in 1995 (fraction) 0.010* (0.006) 0.012** (0.005) Change in share of agriculture GVA in GDP 0.016 (0.011) 0.014 (0.011) Share of agriculture GVA in GDP in 1995 (fraction) 0.016** (0.008) 0.015* (0.008) Country is in South Asia region (ri) -0.036*** (0.014) -0.032** (0.014) ẏi ri 0.473** (0.200) 0.414** (0.198) yi,0 ri -0.014** (0.006) -0.013** (0.006) Constant (α0) -0.003 (0.002) -0.001 (0.002) Coefficient on spatial lag of dependent variable (λ) 0.930*** (0.068) 0.874*** (0.100) Coefficient on spatially autoregressive error (ρ) 0.901*** (0.094) 0.827*** (0.134) Observations 176 176 Pseudo R 2 0.200 0.481 Notes and sources: (se) = standard errors; * significant at 90% level, ** significant at 95% level, *** significant at 99% level. See sources and variable explanations for table 2 above. CLEARING THE AIR: A TALE OF THREE CITIES | 61 Table A.3 provides the estimated income elasticities of ambient PM2.5 for countries at the same level of GDP per capita as India in 1995 (a) and 2015 (b). The estimated elasticities are shown for countries in South Asia compared to countries outside South Asia. They can be interpreted roughly as the percent change in mean annual PM2.5 associated with a one- percent change in GDP per capita. Higher elasticity values suggest that pollution increases by a larger margin for each one-percent increase in income. These elasticities represent the direct effect of income growth on ambient PM2.5 in a country after removing the spillover effects of pollution from neighboring countries. As the table shows, the estimated income elasticities are slightly higher with the ML model, but the differences between South Asia and other regions are large and broadly consistent across all the models. TABLE A.3: Estimated elasticity values of mean annual PM2.5 with respect to GDP per capita for countries at India’s income level in 1995 (a) versus 2015 (b), in and outside the South Asia region In South Asia Outside South Asia (a) (b) (a) (b) .392 .384 .041 .032 GS2SLS model 1 (.189)** (.188)** (.025)* (.025) .374 .360 .040 .027 GS2SLS model 2 (.159)** (.159)** (.027)* (.027) .575 .562 .064 .050 ML model 1 (.219)*** (.219)** (.035)* (.034) .490 .476 0.049 0.035 ML model 2 (.210)** (.210)** (0.034) (0.032) Notes: * significant at 10% level, ** significant at 5% level; (a) India’s GDP per capita in 1995 was US$ 621 and (b) in 2015 was US$ 1,752 (constant year 2010 prices, market rates); “OLS” = (non-spatial) ordinary least squares; “GS2SLS” = generalized spatial two-stage least squares; “ML” = maximum likelihood; the implied elasticity values represent the direct average marginal effect of GDP per capita growth on PM2.5 exposure, separate from spillover effects from neighboring countries. REFERENCES Dasgupta, S., B. Laplante, H. Wang, D. Wheeler. 2002. Confronting the Environmental Kuznets Curve. Journal of Economic Perspectives 16(1) 147-168. Keene, A., and S. C. Deller. 2013. Evidence of the Environmental Kuznets’ Curve among US Counties and the Impact of Social Capital. International Regional Science Review 38(4) Stern, D. I. 2004. The rise and fall of the environmental Kuznets curve. World Development, 32(8), 1419-1439. Stern, D.I. and Van Dijk, J. (2016). Economic growth and global particulate pollutant concentrations. CCEP Working Papers 1604. Wang, Y., W. Chen, Y. Kang, W. Li, F. Guo. 2018. Spatial correlation of factors affecting CO2 emissions at the provincial level in China: A geographically weighted regression approach. 62 | CLEARING THE AIR: A TALE OF THREE CITIES CLEARING THE AIR: A TALE OF THREE CITIES | 63 1818 H Street, NW Washington, D.C. 20433 USA Telephone: 202-473-1000 Internet: www.worldbank.org/environment