The World Bank Economic Review, 31(1), 2017, 271–294 doi: 10.1093/wber/lhv056 Article The Effect of Metro Expansions on Air Pollution in Delhi Deepti Goel and Sonam Gupta Abstract The Delhi Metro (DM) is a mass rapid transit system serving the National Capital Region of India. It is also the world’s first rail project to earn carbon credits under the Clean Development Mechanism of the United Nations for reductions in CO2 emissions. We analyze whether the DM led to localized reduction in three transportation source pollutants. Looking at the period 2004–2006, one of the larger rail extensions of the DM led to a 34 percent reduction in localized CO at a major traffic intersection in the city. Results for NO2 are also suggestive of a decline, while those for PM2:5 are inconclusive due to missing data. These impacts of pollutant reductions are for the short run. A complete accounting of all long run costs and benefits should be done before building capital intensive metro rail projects. JEL classification: Q5, R4 Key words: Air Pollution, Delhi Metro, Urban Transport The Delhi Metro (DM) is an electric-based mass rapid rail transit system mainly serving the Indian National Capital Territory (NCT) of Delhi. The NCT of Delhi covers an area of 1,483 square kilometers and has a population of 16.8 million people according to the Indian Census of 2011, making it one of the world’s most densely populated cities.1 The DM was introduced in 2002 and since then it is being continually extended within the NCT and adjoining areas. As of 2012, its total route length was 190 kilometers, and annual ridership was 0.7 billion (DMRC 2012).2 In this paper we examine whether this Deepti Goel (corresponding author) is an assistant professor in the Department of Economics, Delhi School of Economics, University of Delhi, Delhi 110007; her email is deepti@econdse.org. Sonam Gupta is a research associate with IMPAQ International; her email is sgupta@impaqint.com. The authors are grateful to the journal’s referees for their critical feed- back that substantially improved the paper. They also thank Mr. Amit Kumar Jain from the Delhi Metro Rail Corporation and Dr. D. D. Basu and Mr. D. C. Jakhwal from the Central Pollution Control Board for their help in obtaining some of the data used in the paper. The authors are also grateful to Shantanu Khanna for his excellent research assistance. A supple mental appendix to this article is available at http://wber.oxfordjournals.org/. 1 According to a worldwide ranking of cities by City Mayors Statistics, Delhi ranked thirteenth in terms of population density with 11,050 persons per square kilometer. Mumbai ranked first, and Beijing ranked twelfth, with population densities of 29,650 and 11,500, respectively. The data are compiled from various sources and are the most recent avail- able. Source url: http://www.citymayors.com/statistics/largest-cities-density-125.html. Last accessed on March 31, 2015. 2 As of 2014, the Beijing Subway had a route length of 527 kilometers and annual ridership of 3.4 billion. This informa- tion is accessed from Wikipedia. Source url: http://en.wikipedia.org/wiki/Beijing_Subway. Last accessed on March 31, 2015. C The Author 2015. Published by Oxford University Press on behalf of the International Bank for Reconstruction and Development / THE WORLD BANK. V All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. 272 Goel and Gupta important mode of public transportation has had any impact on air pollution in Delhi. We identify the immediate localized effect of extending the DM rail network on air pollution measured at two different locations within the city: ITO, a major traffic intersection in central Delhi, and Siri Fort, a mainly resi- dential neighborhood in south Delhi. Air pollution is measured in terms of three criteria pollutants, namely, nitrogen dioxide (NO2 ), carbon monoxide (CO), and fine particulate matter (PM2:5 ). An impact study of the DM on air pollution is important for two reasons. First, there is substantial scien- tific evidence on the adverse effects of air pollution on human health. Block et al. (2012) provide a review of epidemiological research that shows the link between air pollution and damage to the central nervous system, which may manifest in the form of decreased cognitive function, low test scores in children, and increased risk of autism and of neurodegenerative diseases such as Parkinson’s and Alzheimer’s. They also cite other studies which show that air pollution causes cardiovascular disease (Brook et al. 2010) and worsens asthma (Auerbach and Hernandez 2012). Turning to recent research in economics, Tanaka (2015) finds that regula- tions to curb pollution from coal-based power plants in China led to 3.29 fewer infant deaths per 1000 live births, amounting to a 20 percent reduction in infant mortality rate.3 Ghosh and Mukherji (2014) examine the effect of ambient air quality on children’s respiratory health in urban India and find that a rise in particu- late matter significantly increased the risk of respiratory ailments. Arceo-Gomez et al. (2012) provide evi- dence from Mexico City that a 1 percent increase in particulate matter (PM10 ) over a year led to a 0.42 percent increase in infant mortality, and the corresponding figure for CO is 0.23 percent. Currie and Walker (2011) find that exposure to vehicular emissions around toll plazas in the northeastern United States increased the likelihood of premature births and also resulted in low birth weight. Some other studies that document the adverse health consequences of air pollution for the United States include Moretti and Neidell (2011); Lleras-Muney (2010); and Currie, Neidell, and Schmieder (2009). The second compelling reason for this study is the extent of air pollution in Delhi. According to the World Health Organization’s (WHO) database, Ambient Air Pollution 2014, Delhi is the most polluted city in the world in terms of PM2:5 levels. In 2013, the annual mean concentration of PM2:5 in Delhi was almost twenty times the guideline value prescribed by the WHO.4 The Central Pollution Control Board (CPCB), the national authority responsible for monitoring and managing air quality in India, finds that pollution in Delhi is positively associated with lung function deficits and with respiratory ailments (CPCB 2008a; CPCB 2008b). Guttikunda and Goel (2013) estimate that particulate matter present in Delhi in 2010 led to prema- ture deaths ranging from 7,350 to 16,200 per year and to 6 million asthma attacks per year. As Delhi contin- ues to grow, population and vehicle densities are bound to increase further, making it all the more important to examine whether the expansion of the DM has had an impact on the city’s air quality. Figures 1A through 1E present the pollution picture at ITO during our study period, 2004 to 2006. Each figure shows the 8- or 24-hour average for a specific pollutant along with the corresponding upper limit prescribed by the CPCB.5 There are some noticeable gaps in each series due to missing observa- tions. In spite of this we see a clear seasonal pattern for nitrogen dioxide (NO2 ), carbon monoxide 3 Greenstone and Hanna (2014) find that environmental regulations in India have been effective in reducing air pollution. However, in contrast to Tanaka (2009), they find an insignificant impact of reduced air pollution on infant mortality. For reasons discussed in their paper, they advise readers to be cautious when interpreting their result on infant mortality. 4 The WHO guideline for annual mean concentration of PM2:5 is 10 lg=m3 , and in 2013, the annual mean level of PM2:5 in Delhi was 198lg=m3 . Notably, Delhi’s PM2:5 level far exceeded that of Beijing which was at 56lg=m3 (Ambient Air Pollution Database 2014, WHO). 5 Appendix table S1.1 in the online supplementary appendix S1 (available at http://wber.oxfordjournals.org/) presents the CPCB limits along with those prescribed by the WHO. WHO (2000) maintains that its limits are to be interpreted as guidelines, and individual countries may have different standards based on prevailing exposure levels, and social, eco- nomic, and cultural considerations. We measure violations vis-a-vis the CPCB limits, as they should be more relevant within the Indian context. The World Bank Economic Review 273 Figure 1A. Nitrogen Dioxide, ITO 2004–2006 Figures 1B. Carbon Monoxide, ITO 2004–2006 Figures 1C. Particulate Matter 2.5, ITO 2004–2006 Figure 1D. Sulphur Dioxide, ITO 2004–2006 Figure 1E. Ozone, ITO 2004–2006 Notes: The dashed vertical lines indicate the dates of Delhi Metro expansion. Data for PM 2.5 only available from November 2004 onwards. Source: Authors’ analysis based on pollutant data from the Central Pollution Control Board. (CO), and particulate matter (PM2:5 ), with their levels being higher in winter (November through January) than in summer (April through June). Further, in the cases of NO2 , CO, and PM2:5 , there were a large number of occurrences when their levels exceeded prescribed limits, while there were fewer 274 Goel and Gupta Figure 1. Continued violations for sulphur dioxide (SO2 ) and ozone (O3 ). During this period, NO2 , CO, and PM2:5 , exceeded limits 85, 48, and 78 percent of the time, respectively, while the corresponding figures for SO2 and O3 were much lower at 3 and 0.1 percent, respectively.6 Given that SO2 and O3 were within permis- sible limits most of the time, our analysis focuses on NO2 , CO, and PM2:5 . Another reason for restricting focus to only these pollutants is that while NO2 , CO, and PM2:5 are mainly generated from transportation sources, SO2 and O3 are not. In one of the first pollution inven- tory studies for Delhi, Gurjar et al. (2004) infer that during their study period (1990–2000), transport sector contributed about 82 percent of nitrogen oxides (NOx ),7 and 86 percent of CO. In another study for Delhi conducted in 2007, NEERI (2010) reports that the contribution of vehicles towards NOx , CO, and particulate matter (PM2:5 and PM10 ) was 18, 58, and 59 percent, respectively. For Delhi in 2010, Guttikunda and Calorie (2012) estimate that 67, 28, and 35 percent of NOx , CO, and PM2:5 , respec- tively, can be attributed to vehicles.8 While there is variation across studies in the exact share of 6 During the same period at Siri Fort, NO2 , CO, SO2 and O3 , exceeded prescribed limits, 13, 26, 0, and 5 percent of the time, respectively. PM2:5 was not recorded at Siri Fort. 7 NOx refers to both nitrogen monoxide (NO) and nitrogen dioxide (NO2 ). 8 The contribution of vehicles towards particulate matter reported in NEERI (2010) and in Guttikunda and Calorie (2012) includes the contribution of road dust as well. The World Bank Economic Review 275 transportation sources in generating these pollutants, all of them report substantial shares. On the other hand, NEERI, and Guttikunda and Calorie, report that vehicular emissions were responsible for 0.3 percent and 3 percent, respectively, of SO2 .9 None of these studies look at O3 . However, it is known that O3 is not directly emitted by motor vehicles, but is created through a complicated non- linear process wherein oxides of nitrogen and volatile organic compounds react together in the pres- ence of sunlight (Sillman, 1999). Thus, of the five pollutants for which we have data, motor vehicles constitute a major and direct source of only three, namely, NO2 , CO, and PM2:5 . To the extent that one of the main channels through which the DM is likely to affect air pollution is through its impact on overall levels of vehicular emissions, we focus our attention on these three pollutants. Moreover, Delhi is a heavily motorized city,10 and the consequent vehicular emissions are a matter of serious concern. Theoretical research from transport economics (Vickery 1969; Mohring 1972) postulates the exis- tence of two counteracting effects on air pollution from introducing a new mode of public transporta- tion. On the one hand, the introduction of the new mode could increase overall economic activity, which could in turn generate new demand for intracity trips. New demand for travel could also be created if the availability of rapid public transport results in a relocation of residents away from the city-center, for example, if real estate is cheaper in the suburbs leading to longer commutes to work. Such demand that did not exist before the new mode was introduced is referred to as the traffic creation effect. If part of the new demand is met by private means of transport, then ceteris paribus this should add to existing levels of vehicular emissions and increase air pollution. On the other hand, with the introduction of a new mode of public transportation, commuters who had earlier relied on private means may now switch to the new mode.11 This substitution away from private to public mode of travel is called the traffic diversion effect. Ceteris paribus, the traffic diversion effect should reduce the overall level of vehicular emissions and consequently reduce air pollution. In reality both effects are likely to operate. We hypothesize that the traffic diversion effect is likely to dominate the traffic creation effect in the short run. This is because the processes involved in creating new demand for travel are likely to unfold slowly and over a longer period of time, while the traffic diversion effect can occur almost immediately after the new mode is introduced. Nonetheless, it is important to verify this empirically. To be able to attribute changes in a pollutant measure to the DM, we use the Regression Discontinuity (RD) approach. As we explain below, due to the presence of sporadic sources of pollution in Delhi (such as spontaneous burning of waste), this approach is not ideal for short periods of analysis. We therefore rely on a three-year study period and argue that these sporadic sources of pollution cancel each other within this timeframe, resulting in reliable estimates. Our analysis reveals that soon after some of the larger extensions of the DM there were significant reductions in at least some transportation source pollutants. Specifically, when we consider our entire study period, 2004–2006, we find that the first extension of the Yellow Line, characterized by the largest surge in metro ridership, resulted in a 34 percent reduction in CO at ITO. Additionally, there is 9 Gurjar et al. (2004) do not report this figure for SO2 . 10 Among the 44 reported million-plus cities in India, Delhi had the largest number of registered motor vehicles during 2011–12 with 7.4 million vehicles (TRW 2013). Goel et al. (2015) estimate that of all registered cars and two-wheelers in Delhi in 2011; 59 percent and 42 percent, respectively, are in-use. If we make a very conservative estimate of vehicles in use and assume that only 42 percent of all registered motor vehicles are in use, we get a figure of 3.1 million for in- use vehicles in Delhi. In 2013, the number of in-use vehicles in New York City was 2 million (Department of Motor Vehicles, New York State website; source url: http://dmv.ny.gov/org/about-dmv/statistical-summaries, last accessed on March 31, 2015). Thus, Delhi has at least 55 percent more in-use vehicles than New York City. 11 According to a report by the Delhi Metro Rail Corporation (DMRC, 2008), the DM has already taken the share of 40,000 vehicles. 276 Goel and Gupta suggestive evidence of a decline in NO2 at ITO due to the introduction of the Blue line. We are unable to say anything conclusive about PM2:5 due to poor quality data on this pollutant. The rest of the paper is organized as follows. In section I we briefly describe the institution of the DM. The empirical strategy is explained in section II and the data sources are listed in section III. Section IV presents our empirical results and section V places them in the context of existing literature on air pollution. Section VI ends with policy recommendations. I. Genesis and Expansion of the Delhi Metro The Delhi Metro Rail Corporation Limited (DMRC) was set up in 1995 by the governments of Delhi and India to take over the construction and subsequent operation of the DM. Construction work for the metro began in 1998. The first commercial run took place on December 25, 2002, between Shahdara and Tis Hazari in north Delhi, marking the beginning of operations. The various stages of expansion of the metro rail network were planned keeping in mind the expected demand for transportation from different localities. The rail lines were first laid in areas with a high pop- ulation density and where it was felt that the metro would benefit the largest number of people. Subsequent extensions were similarly motivated. Table 1 details the phase-wise expansion of the DM rail network from its inception in 2002 to 2006. Six extensions were made during our period of study between 2004 and 2006.12 Figure 2 presents a map of the DM rail network as of December 31, 2006. The map shows the different metro extensions, the two air pollution monitoring stations at ITO and Siri Fort, and the weather station at Safdarjung. II. Empirical Strategy We use Regression Discontinuity (RD) to estimate the causal impact of the DM on pollution.13 The basic idea behind this method is explained here. To get at the causal effect we would have ideally liked to com- pare the levels of pollutants after the metro was extended with their levels, in the same place and at the same time, but in the absence of the metro. However, it is impossible to observe both these scenarios. Therefore, we build the scenario without the metro using observed pollution just before the metro exten- sion. Any sudden change in the levels of pollutants just before and just after the metro extension is attrib- uted to the surge in metro ridership observed at the time of the extension and is interpreted to be the causal effect of the metro extension. It is important to note that this interpretation is correct only if it were true that in the absence of the metro extension, and after accounting for discontinuous changes due to other known factors such as changing weather conditions, there would have been a smooth transition in the lev- els of pollutants over time. Later in this section, we talk about the validity of this identifying assumption. Estimation Equation We measure pollution using data from monitoring stations at two different locations within the city, ITO and Siri Fort.14 Table 2 presents pollution statistics at each location, along with weather conditions at Safdarjung, Delhi. ITO has much higher pollution compared to Siri Fort: Average hourly NO2 and CO at ITO are 3.2 and 1.5 times their respective levels at Siri Fort. This is not surprising given that ITO 12 One reason for not studying the period before 2004 is that we do not have pollution data for it. 13 Lee and Lemieux (2010) provide an excellent exposition of this method. 14 In talking to experts at the CPCB, we were told that a monitoring station measures the quality of ambient air passing by it, and it is not possible to demarcate a precise catchment area for which the quality measure would apply. Given that the monitoring stations at ITO and Siri Fort are approximately 9 kilometers apart, we believe that they each mea- sure air quality in two distinct geographies within the city. Some evidence for this is provided in table 2, which shows that average pollutant levels are very different across the two locations. The World Bank Economic Review 277 Table 1. Phase Wise Extension of the Delhi Metro from Inception to 2006 Extension Segment (stations added) Date of ext. Average daily ridership % Change in average dd-mm-yyyy (000s per day)1 daily ridership2 Extensions before our study period Red Line (introduction) Shahdara-Tis Hazari (5) 25-12-2002 Red Line (first extension) Tis Hazari-Inderlok (4) 04-10-2003 Extensions during our study period, 2004–2006 Red Line (second extension) Inderlok-Rithala (8) 31-03-2004 80.1 46 Yellow Line (introduction) Vishwavidyalaya-Kashmere 20-12-2004 130.8 3 Gate (3) Yellow Line (first extension) Kashmere Gate-Central 03-07-2005 143.2 76 Secretariat (6) Blue Line (introduction) Barakhamba-Dwarka (21) 31-12-2005 270.4 56 Blue Line (first extension) Dwarka- Dwarka Sector 9 (6) 01-04-2006 385.7 5 Blue Line (second extension) Barakhamba-Indraprastha (3) 11-11-2006 451.0 15 Source: Delhi Metro Rail Corporation, DMRC. 1. This refers to the average daily ridership for the month preceding the one in which the extension was introduced. However, for the two extensions carried out on the last day of the month (Red Line second extension and Blue Line introduction), it refers to the average daily ridership for the month in which the extension was introduced. For example, average daily ridership was 80.1 and 130.8 thousand per day in March 2004 and November 2004, respectively. 2. This refers to the month-to-month percentage change in average daily ridership for the month in which the extension was carried out. However for the two extensions carried out on the last day of the month, the change is for the month following the one in which the extenison was carried out. For example, average per- centage daily ridership in April 2004 was 46 percent higher than that in March 2004, while for December 2004 it was 3 percent higher than that in November 2004. is a major traffic intersection, while Siri Fort is a mainly residential area. Ideally, we would have liked to know weather conditions specific to each location. However, we only have hourly weather data for Safdarjung, which is fortunately located between ITO and Siri Fort. We use this as the best available proxy for weather conditions at each location. As the dynamics of pollution are likely to be different across the two locations, and also because they are at different distances from the various line expan- sions, we estimate impacts at each location separately. At each location we estimate the impact of a particular metro extension using a time series of hourly pollutant data lying within a symmetric window around that extension’s opening date. We also ensure that there are no other extensions within this window. Thus, a window is characterized by a location l (ITO or Siri Fort) and an extension m. The RD approach is implemented by estimating the following OLS regression within each window: ylt;m ¼ hl0;m þ hl1;m DMm l;m l;m l;m t þ h2 xt þ h3 PðtÞ þ ut (1) ylt;mis pollutant level (in logs) in hour t, at location l, when studying the effect of extension m. DMm t is the discontinuity dummy for extension m: Within each window it takes the value 1 for all time periods after the extension date, and 0 for periods before it.15 xt ; is the vector of covariates and includes controls for weather;16 for hour of the day,17 day of the week, and interactions between these two; and for public 15 We exclude the 24-hour data pertaining to the day of the extension because we do not know the exact hour when the new line became operational. 16 Controls for weather include current and up to 4-hour lags of temperature, relative humidity, wind speed, and rainfall and quartics of both current and 1-hour lags of these weather variables. 17 Figures S1.1a through S1.1c in the supplementary appendix S1 show how the pollutants at ITO vary by hour of the day and by season. Once again we see that pollution is higher in winter compared to summer. Also, there is substantial intraday variation with peak levels reached between 8 pm and 2 am. Similar patterns are observed at Siri Fort. One plausible reason for why pollution peaks during these night hours could be a citywide ban on the entry and movement of heavy goods vehicles (mostly diesel powered trucks) between 6:00 am and 9:00 pm. The substantial intraday varia- tion in pollution calls for inclusion of hour of the day fixed effects in order to improve the precision of our estimates. In 278 Goel and Gupta Figure 2. Delhi Metro Rail Network as of December 31, 2006 Notes: The two air pollution monitoring stations are at ITO and Siri Fort, and are approximately 9 kilometers apart. The weather station is at Safdarjung. The triangles represent the three coal based thermal power plants located within the National Capital Region of Delhi. Source: Created by the authors using Google Earth and data from the Delhi Metro Rail Corporation Limited. holidays and festivals such as Diwali.18 P(t) is a third-order polynomial in time and captures all smooth variations in pollutant levels. ult;m is the error term. The coefficient hl1;m measures the proportionate change in pollutant level at location l as a result of extension m. It is to be interpreted as the immediate localized (at location l) effect on pollution as a result of that particular extension. Since we expect the traffic creation effect to be negligible in the short run, we do not expect hl1;m to be positive. If there is a strong traffic diversion effect, hl1;m will be negative, otherwise it will be insignificant. Our identification strategy is similar to that used in Chen and Whalley (2012) (henceforth CW).19 CW look at the effect of the introduction of the Taipei Metro (TM) on air quality in Taipei City. While they use the discontinuity arising from the opening of the metro system, we exploit future discontinuities arising from various extensions of the network. Unlike CW, we do not use the first opening of the metro for two reasons. First, we do not have pollution data that dates back to the time when the metro was introduced. Second, even if we had this data, it would be incorrect to use opening ridership discontinuity for Delhi. This is because there was an unprecedented jump in metro ridership when it was first opened, a large part of which was due to joy rides.20 These joy rides are expected to die out as the novelty of the most specifications season-fixed effects are not included because of short window lengths, typically nine weeks. Choice of window lengths is discussed later in this section. 18 Diwali is a Hindu festival that falls in winter, typically in October or November. It spreads over several days and is cel- ebrated with an ostentatious bursting of firecrackers. It has been documented that air pollution in Delhi shoots up dur- ing and immediately following Diwali (CPCB 2012). It is therefore important to control for this source of pollution. 19 Before Chen and Whalley, Davis (2008) used similar identification to estimate the effect of driving restrictions on air pollution in Mexico City. 20 “On the first day itself, about 1.2 million people turned up to experience this modern transport system. As the initial section was designed to handle only 0.2 million commuters, long queues of the eager commuters wishing a ride formed at all the six stations . . . Delhi Metro was forced to issue a public appeal in the newspapers asking commuters to defer joy rides as Metro would be there on a permanent basis” (DMRC 2008). The World Bank Economic Review 279 Table 2. Descriptive Statistics for Pollution and Weather, 2004 to 2006 Mean Std. dev. Min. Max. Missing Obs.1 (%) ITO (in micrograms per cubic meter) NO2 145.5 109.3 0.226 1,140 18 CO 2,389.5 2,328.5 20.6 25,000 14 PM2.52 144.6 131.8 1.69 1,020 42 Siri Fort (in micrograms per cubic meter) NO2 45.0 52.5 0.771 805 27 CO 1,632.0 1,818.7 2.98 20,400 52 Safdarjung Temperature (deg. C) 25.3 7.9 3.1 43.7 0.01 Relative Humidity (%) 61.3 22.0 5 100 0.0 Wind speed (kmph) 4.7 5.8 0 62 0.02 Rainfall (mm) 0.07 1.0 0 56 0.0 Source: Authors’ analysis based on data provided by the Central Pollution Control Board, CPCB and the India Meteorological Department. 1. Missing Obs. refers to percentage of missing observations in the corresponding hourly series. 2. Data for PM2.5 was only recorded at ITO and only from November 2004 onwards. metro fades away. By using discontinuities in ridership that occur a couple of years after the metro first started, we believe that to a large extent we avoid capturing effects arising from one time rides, and the impact that we measure is closer to the steady-state short-term effect. One of the challenges that we faced in estimating equation (1) is the presence of segments of missing observations in each pollutant series. The last column of table 2 shows the share of missing observations.21 The best pollutant series is CO at ITO, for which 14 percent of observations are missing. PM2:5 , which is only recorded at ITO, has 42 percent missing observations. For the RD strategy to be effective, there can- not be too many missing observations around the extension dates. Therefore, to begin with, we restrict our analysis to only those extensions for which there is a symmetric window of at least nine weeks around the date of extension, wherein missing observations in each included week do not exceed 20 percent of the potential observations.22 Then we look at other window lengths, and finally, for those pollutants with rela- tively good data, we analyze the entire series. In order to ensure correct inference in the presence of serial correlation in pollution, in all our specifications we use standard errors clustered at one week.23 Plausibility of Identifying Assumption Identification of the metro effect breaks down if we have not accounted for an event that has a discontin- uous effect on air quality.24 One example is a citywide strike by private bus operators called on the same day as the extension of the metro. If this happens it would be impossible to disentangle the effect of the 21 We were informed by a CPCB official that missing data could be due to many reasons: power cuts, instrument failure, software malfunction when transferring data to storage device, and disruption in telephone. According to the official, none of these reasons are systematically linked to high or low pollution episodes. Later on, in section IV, we examine whether there is evidence for a systematic pattern to missing observations. 22 At the hourly frequency, the number of potential observations in a week is 168 ¼ 24*7. The 20-percent rule implies that each week in our estimation window has at least 134 ¼ 0.8*168 observations. 23 Although, both Chen and Whalley (2012) and Davis (2008) use standard errors clustered at five weeks, we cluster at one week. This is because our analysis is based on shorter windows of five and nine weeks (due to missing data), while they use two-year horizons. Also, for all pollutants in our data, the auto correlation in daily average pollutant level is less than 0.5 beyond seven days. Clustering at one week should therefore be sufficient. Nonetheless, we re-estimated ta- bles 4 and 5 by clustering at two weeks and found similar results. 24 An event that has a gradual effect on pollution will be captured by the time polynomial P(t) and therefore does not im- pede our analysis. 280 Goel and Gupta metro from that of the strike. We have studied the chronology of events in the city and do not find occur- rences of such events on any of the extension dates. Here we discuss some of the other likely threats to identification.25 Government policies aimed at reducing pollution may have an abrupt impact. One such policy, imple- mented only in Delhi, was the mass conversion of diesel fueled buses to compressed natural gas (CNG). However, this happened in 2001, well before our study period began, and is therefore not problematic. In 2005, Delhi moved from Bharat Stage-II to the stricter Bharat Stage-III emission standards. Although this regulatory change was implemented in the middle of our study period, it is unlikely to have led to a sudden change in pollution. This is because the improved norms are only applicable for vehicles manu- factured after the new standards were adopted. Given that new vehicle registrations happen uniformly over time, adoption of stricter emission standards should not lead to a sudden drop in vehicular emis- sions.26 We do not know of any other regulatory change implemented between 2004 and 2006 that may have had a discontinuous effect on pollution. Another concern could be that construction activity undertaken to build the new rail lines may have added to localized pollution in the period preceding the metro extension and this would then over- estimate the DM effect. On speaking to officials from the DMRC we were told that such construction activity is typically completed fifteen to thirty days prior to the opening of a new line so as to conduct trial runs to ensure safety of passengers. Therefore, at least for the shorter window lengths, we do not expect this issue to be a problem. Another worry could be that metro officials choose the extension dates in a systematic manner to coincide with either high- or low-pollution days. We think that this is highly unlikely. Given the public enthusiasm for the metro and the recognition of economies of scale in its oper- ation, the DMRC has always been eager to open a new line once it had met all safety requirements. Finally, Delhi is characterized by a multitude of pollution sources. According to Guttikunda and Calorie (2012), domestic sources, such as burning of biofuel for cooking and heating, use of diesel gener- ator sets, waste burning, and construction, together account for 20, 19, and 26 percent of NOx ; CO, and PM2:5 ; respectively. These sources tend to be sporadic and sometimes mobile, and it is possible that we have not accounted for all of them. In the results section we talk about what definitive conclusions may be drawn in spite of this threat to identification. III. Data Sources All the data used in this study are from secondary sources. Data on pollutants were obtained from the Central Pollution Control Board (CPCB), which collects it as part of the National Air Quality Monitoring Program (NAMP). We use hourly pollution data recorded at two monitoring stations in the city, namely at ITO and at Siri Fort.27 Both are immobile stations that operate on electricity. They pro- vide comparable data as they were bought from the same manufacturer and followed the same 25 For them to be problematic, the discontinuous effects do not have to necessarily happen on the extension date. Discontinuous effects arising anywhere within our short windows would be problematic for estimating the correct causal effect of the metro extension. 26 Emission standards in India are adopted in a phased manner with stricter norms first being implemented in major cities, including in Delhi, and then extended to the rest of the country after a few years. Given that inter-state freight that plies through Delhi continues to follow the more relaxed emission standards, the impact of Bharat Stage-III within Delhi is dampened. 27 Under the NAMP there is one other monitoring station located at the Delhi College of Engineering (DCE) in north Delhi. We do not use data from this station because our identifying assumption is unlikely to hold at this location. Compared to ITO and Siri Fort there are many more erratic sources of pollution at DCE. This is because, (a) it is sur- rounded by Badli, a major industrial township; (b) all along its periphery there are other small scale industrial produc- tion units; (c) during our study period the college building itself was undergoing repair and renovation; and (d) DCE is in a mainly rural part of Delhi where sporadic burning of biomass and wood is widespread. The World Bank Economic Review 281 Figure 3. Change in Average Daily Metro Ridership over Previous Month, 2004–2006 Note: The dashed vertical lines indicate the dates of Delhi Metro expansion. Source: Authors’ analysis based on ridership data from the Delhi Metro Rail Corporation Limited. monitoring protocol throughout our study period. Hourly data on weather conditions at Safdarjung, Delhi, were obtained from The National Data Center of the India Meteorological Department. Our choice of study period (2004–2006) was dictated by the overlapping period for which we had both pollu- tion and weather data. The Delhi Metro Rail Corporation (DMRC) provided us with data on metro ridership. IV. Results Before presenting the impact estimates, we investigate whether the data validate a sudden increase in metro ridership at the time of each extension. Ridership Discontinuities For each month, figure 3 shows the percentage change in average daily ridership on the DM over the pre- vious month.28 The exact magnitudes of change are given in the last column of table 1. Except for the introduction of the Yellow Line and the first extension of the Blue Line, the figure shows a significant rise in average daily ridership for the month (or for the following month)29 of each extension. The absence of a significant rise in ridership for the introduction of the Yellow Line (3 percent increase) may be attributed to the fact that it was the first segment of the north-south corridor and also a short segment (3 additional stations) that connected the university to the existing Red Line at a time when the university was closed for the holiday season. Further, besides the university station, the two other stations on this segment are relatively rich neighborhoods where many people may continue to pre- fer private over public transportation. For the first extension of the Blue Line, the insignificant rise in rid- ership (5 percent increase) may be attributed to the relatively lower population density in south-west 28 Actual daily ridership, instead of average daily ridership in a month, would have been ideal in order to check the sud- den increase in ridership at each extension date. However, this data was not available for our study period. 29 For the second extension of the Red Line and the first extension of the Blue Line, we see the surge in ridership in the month following the one in which the extension took place. This is because these extensions were introduced on the last day of the month, and one would therefore expect average ridership to increase only in the following month. 282 Goel and Gupta Delhi where the extension took place.30 Given the necessity of observing a large surge in ridership in order to identify the DM effect, we exclude these two extensions from our analysis. The largest jump in ridership is seen for the first extension of the Yellow Line (76 percent increase), which connects areas having a high population density (North-East and Central districts) to the hub of government offices in Central Secretariat. A large surge in ridership is also seen for the introduction of the Blue Line (56 percent increase) which is the longest extension among all the extensions considered here. Given this ridership pattern we expect to see larger effects for the first extension of the Yellow Line and the introduction of the Blue Line. We also expect larger effects at ITO than at Siri Fort because of its relative proximity to the line expansions and also because it is a major traffic intersection whereas Siri Fort is mainly residential. Impact Estimates: Short Windows In order to estimate equation (1) using fairly good quality data with fewer missing values, table 3 shows the maximum window length (in weeks) around each extension on applying the at-most-20-percent- missing-data criteria for each included week and also subjecting the selection to a minimum window length of five weeks. As an example, if we restrict ourselves to good quality data, we are able to examine the effect of the second extension of the Blue Line on NO2 at ITO using a maximum window length of only thirteen weeks. Of the four extensions characterized by a significant increase in ridership, we are unable to examine the effects of the second extension of the Red Line because of lots of missing observa- tions around its opening date. For all extensions with at least nine weeks of good quality data,31 figures S1.2a through S1.2h in the supplementary appendix S1, try to visually examine whether there is a break in a pollutant series at the extension date using a nine-week window around the date.32 These plots suggest that there is a drop in pollutant level at the time of each extension, and in some cases this drop is large. Next, we estimate equation (1) to arrive at quantitative estimates. Table 4 shows the results from an estimation of equation (1) using a nine-week symmetric window of good quality data around each extension date. For each location, it shows the percentage change in the pollutant level that may be attributed to a specific metro extension.33 Contrary to our expectations, the first extension of the Yellow Line did not lead to a statistically significant drop in the level of NO2 at ITO, but as expected it resulted in a huge drop of 69 percent in CO at ITO. The introduction of the Blue Line resulted in a 31 percent decrease in the level of NO2 at ITO. Its effect on CO at ITO could not be analyzed because of missing data. We had expected the second extension of the Blue Line to lead to smaller declines, and we find that it did not lead to statistically significant reductions in any of the three pollutants. Turning to the effects at Siri Fort, we were only able to examine the second extension of the Blue Line. Our analysis shows that just as for ITO, this extension did not lead to a statistically sig- nificant decrease in either NO2 or CO at Siri Fort. It is to be noted that even where an effect is not statis- tically significant, its sign is always negative and in some cases the magnitude is not insignificant. 30 Of the nine districts of Delhi, the South-West district had the lowest population density of 4,179 persons per square ki- lometer in 2001. The North-East district had the highest: 29,468 persons per square kilometer (Govt. of NCT of Delhi 2008a). 31 For PM2:5 at ITO we only have a seven-week window of good quality data, and therefore we look at this shorter win- dow of seven weeks for it. 32 The scatter points are daily averages of residuals obtained from a regression of (log) hourly pollutant on all the right- hand side variables in equation (1) except the extension discontinuity, DMm t ; and the time polynomial, P(t). The over- laid curve depicts the fitted residuals from a regression of the scatter points on the extension discontinuity and a third order time polynomial. 33 When calculating the percentage change we apply the correction suggested by Kennedy (1981) in the context of inter- preting the coefficient on a dummy variable in a semilogarithmic equation. The World Bank Economic Review 283 Table 3. Window Length (in Weeks) of Good Quality Data around Each Extension Each included week has no more than 20% missing observations At least five such weeks (symmetric window around extension date) NO2 CO PM2.5 ITO Yellow Line (first extension), Jul 03, 2005 19 41 — Blue Line (introduction), Dec 31, 2005 13 5 5 Blue Line (second extension), Nov 11, 2006 13 13 7 Siri Fort Blue Line (second extension), Nov 11, 2006 13 13 — Source: Authors’ analysis based on data provided by CPCB. Table 5 shows the impacts using a shorter window of five weeks. Compared to the nine-week win- dow, although the magnitude of impact of the first extension of the Yellow Line on NO2 at ITO is larger, it is still not statistically significant. The effect on CO at ITO for this extension has increased to 78 percent. Also for the introduction of the Blue Line, the effect on NO2 at ITO has increased to 55 per- cent. Restricting to a shorter window enables us to study the effects of this extension on CO and on PM2:5 : at ITO it led to a decrease of 56 and 53 percent, respectively. The results for the second extension of the Blue Line at ITO and Siri Fort are similar to those seen in table 4 and continue to remain statisti- cally insignificant. As seen in table 3 there are segments of good quality data that span longer than nine weeks. In table 6 we extend the window beyond nine weeks whenever the data permit us to do so. In most cases there is a decrease in magnitude of impact, and none of the effects are significant now. We provide two plausible explanations for the transitory nature of our impact estimates. One explanation is that some of the sporadic and mobile sources of pollution that characterize Delhi’s pollution inventory get captured when we extend the window, and this masks the impacts for longer time periods. Admittedly, this may also happen for shorter windows, and it could even explain the very large magnitudes for some of the estimates seen in tables 4 and 5.34 As discussed in section II, these sporadic and mobile sources of pollution pose a threat to our identification strategy. However, the fact that when we look at shorter time periods we consistently get negative estimates (in table 4 all esti- mates are negative, and in table 5, all except one (which is close to zero), are negative), makes us believe that some of the larger extensions did reduce specific transportation source pollutants even if the exact magnitudes of reduction may not be those reported in tables 4 and 5. Another explanation for the disappearance of effects in table 6 could be that the traffic diversion effects are indeed transitory and over longer time horizons the DM has no impact on pollution. Duranton and Turner (2011) provide evidence in support of this argument. They find that in cities in the United States, increase in road-building and provision of public transport have no impact on vehicle- kilometers-traveled. They reason that reduced congestion on roads, experienced soon after new roads are built, has a feedback effect that induces existing residents to drive more. If this is true for Delhi, then it is possible that soon after the larger extensions were initiated, the DM diverted private traffic which 34 Additionally, it could also explain some of the inconsistencies in our impact estimates. As pointed out earlier in this sec- tion, we had expected the first extension of the Yellow Line to lead to a reduction in both NO2 and CO. However, while we see a substantial reduction in CO, the effect on NO2 is not statistically significant (see tables 4 and 5). There is also a discrepancy across extensions in the sense that while the magnitude of effects on NO2 and CO at ITO are comparable for the introduction of the Blue Line, for the first extension of the Yellow Line these are quite disproportionate (see table 5). 284 Goel and Gupta Table 4. Effect of Delhi Metro on Air Quality: Nine Week Symmetric Window NO2 CO PM2.51 (Percentage change in level of pollutant) ITO Yellow Line First extension (Jul 03, 2005) À6.6 À69.4*** — Std. error (16.5) (10.5) — Number of observations 1457 1497 — Blue Line introduction (Dec 31, 2005) À30.6** — — Std. Error (9.4) — — Number of observations 1639 — — Blue Line second extension (Nov 11, 2006) À10.4 À13.1 À12.4 Std. error (9.8) (8.2) (16.0) Number of observations 1605 1605 1268 Siri Fort Blue Line second extension (Nov 11, 2006) À25.9 À3.0 — Std. error (24.0) (9.9) — Number of observations 1601 1532 — Source: Authors’ analysis based on data described in the text. 1. Given the importance of analyzing PM2.5 we report results for a shorter window of 7 weeks for it. Missing observations in each included week do not exceed 20 percent. Each estimate is calculated from a separate regression (equation (1)), where the dependent variable is the natural logarithm of hourly pollutant. Controls are third order polynomial in time; hour of the day, weekday and interaction between the two; current and up to 4-hour lags of temperature, humidity, wind speed, and rainfall and quartics of current and 1-hour lags of the same weather variables; and dummy variables for public holidays and festivities such as Diwali. Std. errors are clustered at one week. * indicates significantly different from zero at 10 percent level, ** at 5 percent level and *** at 1 percent level. lowered pollution (as seen in tables 4 and 5 using shorter windows) and also reduced road congestion. These reductions in turn incentivized the remaining drivers to drive more, and may have also added some new drivers, thus wiping out the initial effects on pollution (as seen in table 6) and on road conges- tion. This explanation is along the lines of the traffic creation effect discussed earlier. Unfortunately, our data and empirical strategy do not allow us to discern with surety which of these explanations is true. However, our subsequent analysis using data for the entire study period suggests that the effects may not be transitory. Robustness Checks Here we present some robustness checks for the results seen in table 4. We also present results from a new specification that uses all the data for our study period, ignoring the fact that there are missing observations. For reasons stated below, this is our preferred specification. Varying the Order of the Time Polynomial Following CW, we have used a third order time polynomial. However, in specifications similar to ours, Davis (2008) uses a seventh order polynomial. In order to check that our main results are robust to the choice of polynomial order, we ran the regressions in table 4 using a fourth through seventh order poly- nomial and found similar estimates. The results for the fourth and the seventh order specifications are shown in tables S1.2a and S1.2b in the supplementary appendix S1.35 35 We also tried interacting the time polynomial with the discontinuity dummy, DMm t . We do not report these results as, in all cases, the discontinuity dummy drops out due to multicollinearity in our dataset. The World Bank Economic Review 285 Table 5. Effect of Delhi Metro on Air Quality: Five Week Symmetric Window NO2 CO PM2.5 (Percentage change in level of pollutant) ITO Yellow Line first extension (Jul 03, 2005) À21.4 À77.8*** — Std. error (21.3) (6.7) — Number of observations 748 772 — Blue Line introduction (Dec 31, 2005) À55.2** À56.0* À52.8*** Std. error (7.7) (22.8) (5.3) Number of observations 848 848 840 Blue Line Second extension (Nov 11, 2006) À6.4 À10.1 À15.8 Std. error (5.8) (8.7) (15.8) Number of observations 935 935 932 Siri Fort Blue Line second extension (Nov 11, 2006) À14.2 0.3 — Std. error (10.5) (6.2) — Number of observations 934 871 — Source: Authors’ analysis based on data described in the text. Missing observations in each included week do not exceed 20 percent. Each estimate is calculated from a separate regression (equation (1)), where the dependent variable is the natural logarithm of hourly pollutant. Controls are the same as used in table 4. Std. errors are clustered at one week. * indicates significantly different from zero at 10 percent level, ** at 5 percent level and *** at 1 percent level. Accounting for Persistence in Pollution Next we examine whether controlling for lagged pollution alters the estimated impacts. We estimate the equation shown below (l, m superscripts have been suppressed): yt ¼ a0 þ a1 DMt þ a2 xt þ a3 PðtÞ þ b1 ytÀ1 þ b2 ytÀ2 þ b3 ytÀ3 þ b4 ytÀ4 þ et The results are shown in tables S1.3a and S1.3b in the supplementary appendix S1. As expected, when we account for persistence, the instantaneous impacts are lower.36 The two instances in table 4 where we had seen significant drops in pollution continue to remain significant, although the magnitude of the cumulative effects is lower. Artificial Discontinuities For NO2 and CO, we implemented the standard placebo test of using data only from the pretreatment period (before the extension is made) and introducing artificial discontinuities within this period. In some cases we did find significant effects. In the same vein, we estimated equation (1) using the introduc- tion of the Yellow Line, which is characterized by an insignificant jump in ridership (see earlier discus- sion under “Ridership Discontinuities”). For this expansion, we had expected to find insignificant effects, but instead we found significant declines in both NO2 and CO at Siri Fort. These perverse results could be due to sporadic and mobile sources of pollution in Delhi. While we admit that this is a potential threat to identification, the fact that we consistently get negative effects for all discontinuities in tables 4 and 5 and that, in most cases, these effects are larger when we reduce the time window, provides strong support that DM did reduce pollution. Below, we provide additional evidence to support this claim. 36 The instantaneous impact is a1 , and the cumulative effect is calculated by iteratively substituting for y in the equation shown. The cumulative effect is then given by, ½1 þ b1 þ b2 3 4 2 1 þ b1 þ b1 þ 3ðb1 b2 Þ þ 2b1 b2 þ 2b1 b3 þ b2 þ b2 2 þb3 þ b4 Š à a1 . 286 Goel and Gupta Table 6. Effect of Delhi Metro on Air Quality: Longer than Nine-Week Symmetric Window NO2 CO (Percentage change in level of pollutant) ITO Yellow Line first extension (Jul 03, 2005) À5.5 À18.8 Std. error (11.5) (18.7) No. of obs. (window length in weeks) 3259 (19) 7118 (41) Blue Line introduction (Dec 31, 2005) À3.1 — Std. error (20.5) — No. of obs. (window length in weeks) 2281 (13) — Blue Line second extension (Nov 11, 2006) À2.3 À12.2 Std. error (7.6) (10.5) No. of obs. (window length in weeks) 2397 (13) 2397 (13) Siri Fort Blue Line second extension (Nov 11, 2006) À28.4 9.2 Std. error (29.4) (16.2) No. of obs. (window length in weeks) 2392 (13) 2315 (13) Source: Authors’ analysis based on data described in the text. Missing observations in each included week do not exceed 20 percent. Each estimate is calculated from a separate regression (equation (1)), where the dependent variable is the natural logarithm of hourly pollutant. Controls are the same as used in table 4. Std. errors are clustered at one week. * indicates significantly different from zero at 10 percent level, ** at 5 percent level and *** at 1 percent level. Effects on Nontransportation Source Pollutants Our conjecture is that traffic diversion is the main mechanism through which the DM affects pollution in the short run. We can test this by looking at the effects of various extensions on the two nontranspor- tation source pollutants, namely, sulphur dioxide, SO2 ; and ozone, O3 : If our hypothesis is true, then, we should not find any effects on these pollutants. In fact, CW use this as a robustness check for their results. We first explain why using O3 for such a robustness check is invalid in our context. Then, we implement the check using SO2 and discuss the results. Formation of O3 crucially depends on the ratio of volatile organic compounds (VOC) to NOx (Sillman, 1999). The relationship between O3 , VOC, and NOx is nonlinear, and given that there is evidence that NO2 (and perhaps, VOC, as well), was impacted by the DM, ex-ante it is not possible to know whether or not the DM would affect O3 , and if it does, in what direction. Hence, we do not use O3 . Table S1.4 in supplementary appendix S1 presents the results when we estimate equation (1) using SO2 at ITO as the dependent variable. We examine impacts for the first extension of the Yellow Line and for the introduction of the Blue Line; the two discontinuities for which we found significant declines in transportation source pollutants in table 4.37 We had expected to see no impact but table S1.4 shows large positive effects for both extensions. This may be explained by the presence of three coal based power plants within the National Capital Region of Delhi (marked as triangles in figure 2) during our study period. According to Gurjar et al., 2004, and Guttikunda and Calorie, 2012, 68 and 55 percent, respectively, of SO2 in Delhi is generated by these power plants. Figure S1.3 in the supplementary appendix S1 shows wide variation in the monthly power production of these plants. More importantly, it shows that monthly output was either high or rising when the two extensions were made. Since SO2 emissions are strongly correlated with power produced, there will be corresponding variations in SO2 . 37 In this exercise using SO2 , for both discontinuities, we use a nine week window and once again adopt the criteria of at most 20 percent missing observations for each included week. The World Bank Economic Review 287 Thus, it is possible that it is the coal plants that are behind the significant impacts seen in table S1.4. In any case, that the two metro expansions were accompanied by an increase in SO2 increases the credibil- ity that these expansions led to reductions in some transportations source pollutants (as seen in table 4). Missing Data Patterns Since our analysis so far has been based on examining particular segments of good quality pollutant data, it is important that observations should not be missing systematically. For each pollutant, table S1.5 in supplementary appendix S1, provides time series of share of missing observations in each month between 2004 and 2006. Eye balling the data does not suggest a pattern to missing observations. We examine this more thoroughly in two ways. First we examine whether there is any association between data being missing and the timing of metro expansions. To do this we use data for our entire study period, 2004–2006, and estimate equation (1) by replacing the outcome variable with an indicator variable for whether the pollutant measure is missing. In this specification we include all six discontinuity dummies (representing the six metro expansions undertaken in this period), and also include season fixed effects. Results are presented in tables S1.6a and S1.6b for ITO and Siri Fort, respectively, in supplementary appendix S1. At ITO, for NO2 and PM2:5 there is some evidence that observations are less likely to be missing in summer. However, given that for each pollutant some extensions are positive, some negative, and some insignificant, we conclude that the extensions by themselves are not systematically associated with the pollutant measure being missing. Next we investigate whether observations are more likely to be missing on either high or low pollu- tion days. This is done using the entire series for CO and PM2:5 at ITO. The last column of table 2 shows that CO at ITO has 14 percent missing observations, while the corresponding figure for PM2:5 is 42 per- cent. Given these shares of missing observations, we use a specification similar to equation (1), wherein missing status for PM2:5 is regressed on whether or not CO is above the CPCB prescribed limit of 2000lg=m3 (proxy for high pollution). Other right hand side variables are the same except that we drop the discontinuity dummies and the time polynomial, and once again include season fixed effects. The results are shown in appendix table S1.7 in supplementary appendix S1. The coefficient on CO being high is insignificant suggesting that missing status for PM2:5 is not systematically asso- ciated with high pollution episodes. Impact Estimates using Entire Series In order to get correct magnitudes of impact estimates we still need to address the threat posed by spora- dic and mobile sources of pollution in Delhi’s pollution inventory. To do this we examine the entire ser- ies (between 2004 and 2006) for CO at ITO which has the least share of missing observations (14 percent missing) and estimate the following equation: X M yt ¼ d0 þ d1 xt þ d2 PðtÞ þ ci ðDMiÞt þ et (2) i¼1 The variables are similarly defined as in equation (1), and as above, xt now includes season fixed effects. Because we look at the entire series we include all extensions shown in table 3 for ITO and this is repre- sented by the set of discontinuity dummies, fDMig. We expect that the sporadic sources of pollution are likely to be evenly spread within a long window of three years and would therefore cancel each other. This is therefore our preferred specification. Figure S1.4 in the supplementary appendix S1 (similarly constructed as figures S1.2a through S1.2h), visually presents the effects of multiple extensions on CO at ITO. The impact magnitudes are presented in table 7 which shows that the first extension of the Yellow Line, characterized by the largest increase 288 Goel and Gupta Table 7. Effect of Delhi Metro on CO at ITO: Entire Series, 2004–2006 Metro Extension CO at ITO Yellow Line First Extension (Jul 03, 2005) À33.5** Std. Error (10.4) Blue Line Introduction (Dec 31, 2005) À16.3 Std. Error (16.6) Blue Line Second Extension (Nov 11, 2006) À2.9 Std. Error (14.6) Number of Observations 22,158 Source: Authors’ analysis based on data described in the text. The dependent variable is the natural logarithm of hourly CO at ITO. Control variables are the same as used in table 4, with the addition of season fixed effects. Std. errors are clustered at one week. * indicates significantly different from zero at 10 percent level, ** at 5 percent level, and *** at 1 percent level. in ridership, led to a 33.5 percent reduction in CO at ITO, while the other two extensions did not lead to statistically significant reductions. Again, it is reassuring that all point estimates are negative even if some are not statistically significant.38 We also carry out similar analysis for NO2 at ITO, which has 18 percent missing observations for the entire series. Figure S1.5a in supplementary appendix S1 (counterpart of figure S1.4 for CO), seems to suggest that the first extension of the Yellow Line and the second extension of the Blue Line led to an increase in NO2 at ITO. However, looking at the plot, it seems that, unlike CO, the simple time trend does not fully capture the systematic changes for NO2 , and therefore we interact it with the discontinuity dummies. Figure S1.5b presents the new picture. The effects seen in the previous plot disappear. If any- thing, the introduction of the Blue Line seems to have decreased NO2 at ITO. When we estimate equa- tion (2) for NO2 at ITO including the interaction of the discontinuity dummies with the time polynomial, the discontinuity dummy for the second extension of the Blue Line drops out, perhaps due to multicollinearity in our dataset. We present the results in table 8. The coefficients on the discontinuity dummies are not directly comparable to the coefficients in tables 4 through 7 because here we have inter- acted the discontinuity dummies with the time polynomial. We note that the coefficient on the introduc- tion of the Blue Line is negative and significant.39 Evidence for Traffic Diversion Table 9 shows the growth rates in registered private vehicles in Delhi and Mumbai over two five year periods, 1997 to 2002, and 2002 to 2007. In 1997–2002 neither city had a metro system. In the later period the DM was operational in Delhi, while Mumbai did not see any major change in its public transport infrastructure. For each city, we calculate the difference in growth rates between the two time periods. Compared to 1997–2002, Delhi showed a decrease in the growth rate of Cars and Jeeps in the post-metro period, while Mumbai showed an increase. For two-wheelers, both cities showed an increase in growth rates in 2002–2007 compared to the earlier period, but the increase was smaller for Delhi 38 To check whether the result is robust to choice of polynomial order, we ran the regression using fourth through seventh order polynomial in time. The main result does not change: the first extension of the Yellow Line is the only one that is significant, with estimate values of À32.9 and À39.3 for the fourth and seventh order polynomial, respectively. We also estimate equation (2) using the entire series for SO2 at ITO which has 16 percent missing observations. The only significant discontinuity was the introduction of the Blue Line for which the sign was positive. Thus, the result in table 7 is also robust to the check using nontransportation source pollutant SO2 . 39 This reinforces the result in table 4, in which the introduction of the Blue Line was the only extension that showed a significant impact on NO2 at ITO. The World Bank Economic Review 289 Table 8. Effect of Delhi Metro on NO2 at ITO: Entire Series, 2004–2006 Coeff. Std. err. Yellow Line first extension 40.4 99.1 Blue Line introduction À404.1** 161.6 t À0.0002 0.0002 t squared 2.81E-08 3.61E-08 t cubed À1.52E-12 1.78E-12 t*YL1E À0.0085 0.0195 t squared*YL1E 5.71E-07 1.27E-06 t cubed*YL1E À1.19E-11 2.72E-11 t*BLI 0.0623** 0.0278 t squared*BLI À3.21E-06* 1.63E-06 t cubed*BLI 5.53E-11* 3.29E-11 Number of observations 20646 Source: Authors’ analysis based on data described in the text. The dependent variable is the natural logarithm of hourly NO2 at ITO. Control variables are the same as used in table 4, with the addition of season fixed effects and interactions between the discontinuity dummies and the time polynomial. Std. errors are clustered at one week. * indicates significantly different from zero at 10 percent level, ** at 5 percent level, and *** at 1 percent level. Table 9. Growth Rate in Registered Private Vehicles Cars and Jeeps Two-wheelers Delhi Mumbai Delhi Mumbai 1997–2002 (a) 59.4 25.0 25.5 44.5 2002–2007 (b) 43.5 31.8 43.4 66.7 City difference (b) - (a) À15.9 6.9 17.9 22.2 Delhi difference - Mumbai difference À22.7 À4.3 — — Source: Various issues of Road Transport Yearbook, published by the Transport Research Wing, The Ministry of Road Transport and Highways. Each figure in the table is a growth rate over a five-year period expressed in percent terms. compared to Mumbai. This is suggestive of a traffic diversion effect being operative in Delhi after the metro was introduced. V. Discussion of Results We start by summarizing our main results and then discuss our estimates in the context of other studies, especially the one by CW. Summary of Findings When we look at a short span of nine weeks, we find that the first extension of the Yellow Line, charac- terized by the largest surge in ridership, led to a 69 percent reduction in CO at ITO. When we extend this span to forty one weeks, the effect size reduces and is no longer statistically significant. However, when we consider our entire study period, 2004–2006, we find that this extension resulted in a 34 per- cent decline in CO at ITO. The fact that we find a decline when we look at the whole series (our pre- ferred specification) suggests that the effect is not transitory. The introduction of the Blue Line, the longest extension considered here, led to a 31 percent reduction in NO2 at ITO when we look at a nine week window, and the effect remains when we consider the 290 Goel and Gupta entire study period, which once again suggests that this is not a transient effect. Using a five week win- dow, there is some evidence that the introduction of the Blue Line also led to a decline in PM2:5 at ITO, but we could not carry out the analysis for the entire study period due to a large number of missing observations. Finally, we do not find any significant effects at Siri Fort which is mainly a residential area, and relative to ITO, is further away from the extensions considered here. In supplementary appendix S2, we compare our estimate of a 34 percent reduction in CO with the one in Doll and Balaban, 2013. Doll and Balaban also study the effect of the DM on air pollution, but they use a different methodology to do so. The comparison reveals that our impact estimate is many times higher than theirs. We refrain from commenting on the reasons for this difference beyond what we have stated in this appendix. Comparison with Chen and Whalley (2012) CW estimate the impact of introduction of the Taipei Metro (TM), in 1996, on pollution in Taipei City. Using a two-year window of very good quality data (they had 1 missing observation in a two year win- dow), they find that the opening of the TM resulted in a 15 percent decline in CO. Using a three year window and spanning multiple extensions, we find a much larger impact of 34 percent for Delhi.40 One reason for why our impact is larger could be that CO measurements in our study are from a single moni- toring station located at a major traffic intersection, whereas CW use the average CO across ten moni- toring stations and they exclude the few stations located at traffic intersections (see footnote 20 on page 15 in CW). If traffic diversion is the main mechanism via which the metro impacts pollution, then one might expect to see bigger effects at traffic intersections. Next, we compare the implied ridership-pollution elasticities across the two studies. From not having any metro in the city, CW report an average daily ridership of 40,410 in year following the TM intro- duction. Using data from the DMRC, we note that average daily ridership before and after the first extension of the Yellow Line was 119,855 and 385,866, respectively. Thus, a one percentage point decline in CO is associated with an increase in average daily ridership of 2,694 in case of the TM, and an increase of 7,824 for the DM. Why should DM have a higher ridership-pollution elasticity com- pared to the TM? Note that the pre-metro (1995–96) CO level in Taipei was 1; 030lg=m3 , while the CO level for Delhi before the first extension of the Yellow Line was 2; 212lg=m3 .41 Given that base- line pollution in Delhi is about twice that of Taipei, a one percentage point change in pollution would translate into a much larger absolute reduction in pollution for Delhi. Therefore, it is to be expected that Delhi would require a larger change in ridership to support a larger absolute change in pollution. Moreover, compared to Taipei, a much larger proportion of Delhi’s population uses public trans- port.42 For traffic diversion from private vehicles to take effect, Delhi would therefore need a much larger increase in ridership. 40 Taipei today has a population density of about 15,200 persons per square kilometre, slightly higher than Delhi’s 11,050 (Data on population densities were accessed from City Mayors Statistics. Source url: http://www.citymayors. com/statistics/largest-cities-density-125.html. Last accessed on March 31, 2015). According to an inventory study for Taipei City (Chang and Lee, 2006, as cited in CW), 96 percent of CO was due to vehicles. This is comparable to the 86 percent for Delhi reported in Gurjar, 2004, but is much higher than the 58 percent reported in NEERI, 2010. 41 CW present pollution levels in parts per million (ppm). We converted their figures into micro grams per cubic meter ðlg=m3 Þ using an online convertor provided by Lenntech. Source url: http://www.lenntech.com/calculators/ppm/ converter-parts-per-million.htm. Last accessed on March 31, 2015. 42 The travel mode shares for Taipei City in 2001 (five years after the opening of TM), were 8.8 percent TM, 16.1 percent bus, 34 percent car and 41.1 percent motorbike (Jou et al. as cited in CW). The travel mode shares for Delhi in 2007 (also 5 years since DM opening), were 21 percent walk, 12 percent cycle, 5 percent two-wheelers, 43 percent public transport, 14 percent car and 6 percent three wheelers (Ministry of Urban Development, 2008). The World Bank Economic Review 291 Benefits from Reduced CO Pollution Drawing from literature that looks at health impacts of air pollution, we quantify the benefits of reduced CO levels in Delhi in terms of infant lives saved. Currie, Neidell, and Schmieder (2009) study the effect of pollution on infant health in New Jersey in the 1990s. They account for avoidance behavior, and esti- mate that a 1 part per million (ppm) reduction in CO, reduced infant mortality by 2.5 percent in propor- tion to the baseline risk. CW combine this with their estimate of a 15 percent reduction in CO, and conclude that the opening of the TM resulted in 1.7 infant lives being saved in the year following the metro opening. We undertake a similar exercise and find that the expansion of the DM (first extension of the Yellow Line), resulted in about 61 infant lives being saved in the year following the extension.43 Viscusi (2008) argues that less developed countries tend to have a lower value of statistical life (VSL), consistent with their lower levels of income. He calculates a VSL of 1.2–1.5 million USD (2000 prices) for India, and 0.2–0.9 for Taiwan, as against 7 for the U.S. Using the lower bounds of VSL estimates for India and Taiwan the infant lives saved may be valued at 73.2 and 0.34 million USD (2000 prices), respectively. Given that baseline pollution levels in Delhi are higher than in Taipei, and perhaps, avoid- ance behavior is much less in Delhi due to lack of pollution related warnings, we conjecture that this figure is a lower bound for Delhi. How Viable is the Delhi Metro? Winston and Maheshri (2007) estimate the contribution of each U.S. urban rail operation to social wel- fare. They find that with the exception of the BART, the San Francisco Bay area metro system, every sys- tem actually reduced welfare. They reason that rail systems are unviable for most U.S. cities because of their high capital costs, declining demand for rail travel, rising labor costs, and inability to raise fares as they have to compete with bus services. The authors report that, on average, rail transit systems in the U.S. cover only about 40 percent of their operating costs. This should raise concerns about the viability of the DM, pollution benefits notwithstanding. Supplementary appendix table S1.8 shows the annual profits for the Delhi Metro Rail Corporation Ltd. from 2004 to 2010. It is heartening to note that the company recorded positive profits from traffic operations throughout this period and positive profits from all operations in most years. Delhi has a growing population and high rates of economic growth, and this should alleviate concerns around low ridership demand.44 As of 2012, the total route length of the DM was 190 kilometers, and its ridership was about 2 million per day (DMRC 2012). In contrast, the San Francisco/Oakland area, has a popula- tion density that is slightly more than one fourth that of Delhi’s, and in 2014, the BART’s route length and average daily ridership was 167 kilometers and 0.4 million, respectively.45 Moreover, Delhi is a pol- ycentric city and the route design of the DM is ideal to serve daily commuters. It has a radial track lay- out, with major north-south and east-west corridors connecting different parts of Delhi to the border cities such as Gurgaon and Noida, which are the new centers of employment. Also, ridership is likely to increase further as new routes are completed. According to Litman (2014) rail transit is more appropriate 43 For Delhi in 2004, on average, there were 838 births per day and the infant mortality rate (IMR) was 13 per 1000 in- fant population (Govt. of NCT of Delhi. 2008b). The pre-extension CO level was 2212 lg=m3 ¼ 1:794ppm. Given our impact estimate of 34 percent and a 2.5 percent IMR improvement estimate, annual infant lives saved is calculated as: 1.794*0.34*0.025*0.013*(305870)¼61. The corresponding estimate for Taipei (as shown in CW) is: 0.834*0.156*0.025*0.00666*(77029)¼1.7. 44 The decadal growth in its population from 2001 to 2011 was 20.96 percent, and for the period 2007–2011, the annual compound growth of its Gross State Domestic Product was 10.1 percent (Govt. of NCT of Delhi 2012). 45 Data on population density for San Francisco was accessed from City Mayors Statistics. Source url: http://www. citymayors.com/statistics/largest-cities-density-125.html. Route length and daily ridership on the BART was from http://www.bart.gov/sites/default/files/docs/2014BARTFactsheet_Final%20011614.pdf. Both sites were last accessed on March 31, 2015. 292 Goel and Gupta in areas where development is more compact and noise and air pollution are serious considerations, while buses are more appropriate where travel is more dispersed. It seems to us that Delhi meets the criteria that favor having a metro. For these reasons we are optimistic about the viability of the DM.46 VI. Concluding Remarks We study the impact of the Delhi Metro (DM) on pollution in Delhi between 2004 and 2006. We find that soon after the first extension of the Yellow Line of the DM, there was a significant reduction in car- bon monoxide at a major traffic intersection in central Delhi. There is also suggestive evidence that the introduction of the Blue Line led to a decline in nitrogen dioxide at the same intersection. Although we find a favorable impact of the DM on Delhi’s pollution soon after some of the larger metro expansions, the overall impact of the metro system on air pollution crucially depends on how the electricity needed to drive the metro is generated. If it is not cleanly generated, then some or all of the benefits from reduced pollution in Delhi may be offset by increased pollution elsewhere.47 Also, our impact estimates are for the short run, and quantifying the net social benefit of the metro network in the long run is beyond the scope of our analysis. Much of our analysis was restricted by the poor quality of data on pollution. A longer time series, with fewer missing observations, would have allowed us to draw more definitive conclusions. Given the severity of Delhi’s pollution problem, we would urge the environment ministries at the state and the cen- ter to invest in better technology and equipment in order to record pollution levels more accurately and completely. As our analysis reveals, it is difficult to conduct rigorous impact assessments without good quality data. Our paper also highlighted the nature of pollution in Delhi by citing several inventory studies. We believe that a multipronged strategy needs to be adopted if Delhi wants to shed its distinction of being the most polluted city in the world (based on WHO’s Pollution Database: Ambient Air Pollution 2014). As potential pollution abatement measures, Delhi’s planners should consider the following: increasing accessibility to metro stations by improving feeder systems, promoting cycle-rickshaws to deliver last mile connections, extending the Delhi Bus Rapid Transit System (BRTS) and ensuring strict enforcement of existing bus corridors, using congestion pricing and citywide parking policies to dissuade use of pri- vate vehicles, adopting uniform emissions standards throughout the country, constructing a bypass road around Delhi to eliminate interstate freight traffic, designing efficient waste disposal systems to prevent sporadic burning of garbage and foliage, shutting down the remaining coal based power plants in the city, and educating Delhi’s residents of the severity of the problem to enable them to take more informed decisions. 46 Murty et al. (2006) carry out a social cost benefit analysis of phases I and II of the DM, covering a track length of 108 kilometers. Considering the estimates of financial flows during the period 1995–2041, they estimate the financial benefit-cost ratio to be between 1.92 and 2.30. They also estimate the capital costs of phases I and II to be 64,060 mil- lion and 80,260 million, respectively, and the net present social benefit from both phases to be 419.98 billion INR (in 2004–05 prices). Their calculations account for differences in shadow and market prices of unskilled labor; premiums for importing fuel; benefits accruing from reduced road congestion, accidents, and air pollution; and effects of redistri- bution of income among stakeholders. Their estimates are based on several assumptions regarding annual flows of costs and benefits during the entire lifetime of the DM. Evaluating the accuracy of these assumptions is beyond the scope of our study. 47 The DM has a regenerative braking system on its rolling stock, which generates electricity when brakes are applied and then feeds it back into the system. This way, almost 35 percent of the electricity consumed is regenerated by the system (Sreedharan 2009). Doll and Balaban (2013) provide an excellent analysis of the overall carbon footprint of the DM. For the year 2011, they estimate that the DM saved 232,162 tons of CO2 because of its regenerative braking technology. The World Bank Economic Review 293 Finally, before investing in a capital intensive rail network, it is imperative that a proper cost benefit analysis be undertaken. A specific consideration that needs to be made is its desirability vis- a-vis a bus transit system (BTS). In the context of the United States, Winston and Maheshri (2007) and O’Toole (2010) claim that the less capital intensive BTS is more suitable for most US cities where there isn’t enough demand for the metro to be able to recoup its high capital costs. Litman (2014) makes conten- tions in favour of the metro, stating that while it is more capital intensive, it has lower operating costs per passenger mile. Both sides, however, agree that metro systems are best suited in areas with high pop- ulation density, characterized by high demand for travel. Today, governments at various levels in India are planning to build metro rail systems in several tier II cities. These cities have lower population den- sities compared to tier I cities such as Delhi. Governments, at times, invest in capital intensive projects like the metro to emulate more advanced economies or to pander to private firms who build and design these systems. We caution them against these temptations. As many cities in the United States are realiz- ing, once a metro is built, it is hard to abandon it (Winston and Maheshri 2007). References Arceo-Gomez, E. O., R. Hanna, and P. Oliva. 2012. “Does the Effect of Pollution on Infant Mortality Differ between Developing and Developed Countries? Evidence from Mexico City.” National Bureau of Economic Research, Cambridge, MA; Working Paper 18349. Auerbach, A., and M. L. Hernandez. 2012. “The Effect of Environmental Oxidative Stress on Airway Inflammation.” Current Opinion in Allergy and Clinical Immunology 12 (2): 133–39. Block, M. L., A. Elder, R. L. Auten, S. D. Bilbo, H. Chen, J. C. Chen, and D. A. Cory-Slechta, et al. 2012. “The Outdoor Air Pollution and Brain Health Workshop.” NeuroToxicology 33 (5): 972–84. Brook, R. D., S. Rajagopalan, A. C. Pope, III, J. R. Brook, A. Bhatnagar, A. V. Diez-Roux, and F. Holguin et alet al. 2010. “Particulate Matter Air Pollution and Cardiovascular Disease: An Update to the Scientific Statement From the American Heart Association.” Circulation 121: 2331–2378. Chen, Y., and A. Whalley. 2012. “Green Infrastructure: The Effects of Urban Rail Transit on Air Quality.” American Economic Journal: Economic Policy 4 (1): 58–97. CPCB. 2008a. “Epidemiological Study on Effect of Air Pollution on Human Health (Adults) Delhi.” Central Pollution Control Board, Delhi; Environmental Health Series 1. ———. 2008b. “Study on Ambient Air Quality, Respiratory Symptoms and Lung Function of Children in Delhi.” Central Pollution Control Board, Delhi; Environmental Health Series 2. ———. 2012. “Ambient Air and Noise Pollution Levels - Deepawali 2012.” Central Pollution Control Board, Delhi; Press Release, November 13, 2012. Currie, J., M. J. Neidell, and J. F. Schmieder. 2009. “Air Pollution and Infant Health: Lessons from New Jersey” Journal of Health Economics 28 (3): 688–703. Currie, J., and R. Walker. 2011. “Traffic Congestion and Infant Health: Evidence from E-Z Pass.” American Economic Journal: Applied Economics 3 (1): 65–90. Davis, L. W. 2008. “The Effect of Driving Restrictions on Air Quality in Mexico City.” Journal of Political Economy 116 (1): 38–81. DMRC. 2008. “A Journey to Remember.” Public Relations Department, Delhi Metro Rail Corporation Limited, New Delhi. ———. 2012. “Annual Report 2011–12.” Delhi Metro Rail Corporation Limited, New Delhi. Doll, N.H.C., and O. Balaban. 2013. “A Methodology for Evaluating Environmental Co-Benefits in the Transport Sector: Application to the Delhi Metro” Journal of Cleaner Production 58 (1): 61–73. Duranton, G., and M. A. Turner. 2011. “The Fundamental Law of Road Congestion: Evidence from US Cities.” American Economic Review 101 (6): 2616–52. Ghosh, A., and A. Mukherji. 2014. “Air Pollution and Respiratory Ailments among Children in Urban India: Exploring Causality.” Economic Development and Cultural Change 63 (1): 191–222. 294 Goel and Gupta Goel, R., S. K. Guttikunda, D. Mohan, and G. Tiwari. 2015. “Benchmarking Vehicle and Passenger Travel Characteristics in Delhi for On-road Emissions Analysis.” Travel Behaviour and Society 2: 88–101. Govt. of NCT of Delhi. 2008a. “Economic Survey of Delhi 2008–2009.” Planning Department, Government of NCT of Delhi. ———. 2008b. “Statistical Abstract 2008.” Directorate of Economics and Statistics, Government of NCT of Delhi. ———. 2012. “Economic Survey of Delhi 2012–2013.” Planning Department, Government of NCT of Delhi. Greenstone, M., and R. Hanna. 2014. “Environmental Regulations, Air and Water Pollution, and Infant Mortality in India.” American Economic Review 104 (10): 3038–72. Gurjar, B. R., J. A. Aardenne, J. Lelievelda, and M. Mohan. 2004. “Emission Estimates and Trends (1990–2000) for Megacity Delhi and Implications.” Atmospheric Environment 38 (33): 5663–81. Guttikunda, S., and R. Goel. 2013. “Health Impacts of Particulate Pollution in a Megacity- Delhi, India.” Environmental Development 6 (1): 8–20. Guttikunda, S., and G. Calori. 2012. “Multi-Pollutant Emissions Inventory for the National Capital Region of Delhi.” SIM-air Working Paper Series, New Delhi, No. 38. Kennedy, P. E. 1980. “Estimation with Correctly Interpreted Dummy Variables in Semilogarithmic Equations.” American Economic Review 71 (4), 801. Lee, D., and T. Lemieux. 2010. “Regression Discontinuity Designs in Economics.” Journal of Economic Literature 48: 281–355. Litman, T. 2014. “Evaluating Rail Transit Criticism.” Victoria Transport Policy Institute, Victoria, BC. Lleras-Muney, A. 2010. “The Needs of the Army: Using Compulsory Relocation in the Military to Estimate the Effect of Air Pollutants on Children’s Health.” Journal of Human Resources 45 (3): 549–90. Ministry of Urban Development. 2008. “Study on Traffic and Transportation Policies and Strategies in Urban Areas in India: Final Report.” Ministry of Urban Development, Delhi. Moretti, E., and M. J. Neidell. 2011. “Pollution, Health, and Avoidance Behavior: Evidence from the Ports of Los Angeles.” Journal of Human Resources 46 (1): 154–75. Mohring, H. 1972. “Optimization and Scale Economies in Urban Bus Transportation.” American Economic Review 62 (4): 591–604. Murty, M. N., K. K. Dhavala, M. Ghosh, and R. Singh. 2006. “Social Cost-Benefit Analysis of Delhi Metro.” Institute of Economic Growth, Delhi; Munich Personal RePEc Paper No. 14465. NEERI. 2010. “Air Quality Monitoring, Emission Inventory, and Source Apportionment Studies for Delhi.” National Environmental Engineering Research Institute, Nagpur; December 2010. O’Toole, R. 2010. “Defining Success: The Case Against Rail Transit.” Policy Analysis 663: 1–40. Sillman, S. 1999. “The Relation between Ozone, NOx and Hydrocarbons in Urban and Polluted Rural Environments.” Atmospheric Environment 33: 1821–1845. Sreedharan, E. 2009. “Climate Change—An Opportunity for Sustainable Development: The DMRC Experience.” RITES Journal: 81–88. Tanaka, S. 2015. “Environmental Regulations on Air Pollution in China and their Impact on Infant Mortality.” Journal of Health Economics 42: 90–103. TRW. 2013. “Road Transport Year Book (2011–12).” Ministry of Road Transport and Highways, Government of India, Transport Research Wing, New Delhi. Vickery, W. 1969. “Congestion Theory and Transport Investment.” American Economic Review 59 (2): 251–60. Viscusi, W. K. 2008. “Value of Life.” In S. N. Durlauf, and L. E. Blume, eds., The New Palgrave Dictionary of Economics. Second Edition. London: Palgrave Macmillan. WHO. 2000. “Air Quality Guidelines for Europe.” World Health Organization, 2nd Edition, WHO Regional Publications, Copenhagen; European Series, No. 91. ———. 2006. “Air Quality Guidelines for Particulate Matter, Ozone, Nitrogen dioxide and Sulphur dioxide. Global Update 2005.” World Health Organization, Regional Office, Copenhagen. Winston, C., and V. Maheshri V. 2007. “On the Social Desirability of Urban Rail Transit Systems.” Journal of Urban Economics 62: 362–82.