







































 

Ecology, Economy and Society–the INSEE Journal 2 (1): 93–125, January 2019 

 
RESEARCH PAPER 
 

Quantifying Air Pollution Vulnerability and its Distri-
butional Consequences: some perspectives from Delhi 
 

Shivani Gupta, Sukanya Das  and M.N. Murty  
 
Abstract: This paper estimates Vulnerability Index of air pollution in Delhi taking 
into account exposure, susceptibility and coping capacity of households. A general 
health production function model and a vulnerability assessment framework are 
used for this purpose.  Data was collected through a survey of sample households 
located in close vicinity to 10 air pollution monitoring stations in Delhi. The 
estimated vulnerability index is used to show the effect of household exposure to 
air pollution. The vulnerability index takes into consideration sample households’ 
socio-economic status, demographic profile and other characteristics. Result 
showed that households of lower socio-economic status were the most vulnerable 
to air pollution and its consequences.  

The study also quantifies the economic benefits to Delhi households from 
reduction in air pollution to the standard safety limits of PM10 (100 µg/m3). 
Estimates show that the total annual economic (health) benefits for a typical 
household is Rs. 33,978 and for the whole population of Delhi is Rs. 52.4 billion. 
The study also found that a household of a lower socio-economic status could save 
much more out of their annual income (4.96 per cent) as compared to a household 
of a higher socio-economic status (1.97 per cent) from reduced air pollution. 

Keywords: Environmental Inequity; Air Pollution; Health Production Function; 
Vulnerability Framework, Coping Capacity; Health Benefits 

 
 Research Analyst, Poverty, Health and Nutrition Division, International Food Policy 

Research Institute (IFPRI), New Delhi 110008, India; shivaniguptaa76@gmail.com.  
 Associate Professor, Department of Policy Studies. TERI School of Advanced Studies,10, 
Institutional Area, Vasant Kunj, New Delhi 110070, India; sukanya.das@terisas.ac.in. 
 Former Professor, TERI School of Advanced Studies,10, Institutional Area, Vasant 
Kunj, India, New Delhi 110070 and Former Professor, Institute of Economic Growth, New 

Delhi 110007, India; mn.murty71@gmail.com. 

This study was carried out in 2015-16 as a part of Masters’ Thesis of the corresponding 
author at TERI School of Advanced Studies, New Delhi.   

Copyright © Gupta, Das and Murty 2019. Released under Creative Commons Attribution-
NonCommercial 4.0 International licence (CC BY-NC 4.0) by the author.  

Published by Indian Society for Ecological Economics (INSEE), c/o Institute of Economic 
Growth, University Enclave, North Campus, Delhi 110007.  

ISSN: 2581-6152 (print); 2581-6101 (web). 

DOI: https://doi.org/10.37773/ees.v2i1.57  

https://doi.org/10.37773/ees.v2i1.57


Ecology, Economy and Society–the INSEE Journal [94] 

 

Figure 1: PM10 levels (Specific cities, period 2008-2012) 

 

Source: Figure 3 in WHO (2014, 2)  

Abbreviations: PM10: Fine particulate matter of 10 microns or less; Afr: 
Africa; Amr: America; Emr: Eastern Mediterranean; Eur: Europe; Sear: 

South‐East Asia; Wpr: Western Pacific; LMI: Low‐ and middle‐income; HI: 

high‐income. 

1. INTRODUCTION 

1.1. Background  

Air pollution has very significant health effects on urban households in 
major cities of India, especially in the national capital of Delhi. WHO 
(World Health Organization) (2014) report found that Delhi had the most 
astounding mean level of particulate air pollution among 1,600 noteworthy 
urban cities around the world. Delhi had PM10 concentration of 286 µg/m3 
compared to Beijing’s 121 µg/m3 during the period 2008-2012. This is four 
times more than the permissible limits of PM10 (100 µg/m3) as shown in 
Figure 1. Recently, a report (World Bank 2015) in its study of 381 cities 
from developing countries globally branded Delhi as the worst in air 
pollution.  

Amongst the many causes of air pollution in Delhi, industrial and vehicular 
pollution remain as major factors for emissions of PM10. They emit PM 
(particulate matter) along with COx, NOx, and SOx (carbon, nitrogen and 
sulfur oxides) and metal compounds in the air. A judgment of the Supreme 
Court of India in 2001, in this regard, compelled all Delhi public transport 



[95] Shivani Gupta, Sukanya Das and M.N. Murty 

 

to run on Compressed Natural Gas (CNG), which led to a reduction in 
PM10 levels. Similarly, the Delhi metro operation also helped to reduce 
levels of pollution. However, over the years, these gains are seemingly 
nullified by rapid urbanization and growing vehicular pollution. In 2016, a 
15-day “even-odd policy” was also rolled out by the government to ration 
road traffic and its pollution. Even though this was a good experiment for 
curbing vehicular air pollution, it was not altogether successful in producing 
significant results, leaving Delhi still polluted as it was earlier. Previous 
studies have shown that the health (morbidity and mortality) effects of air 
pollution are severe in key urban areas of both developed as well as 
developing countries. However, one important problem is to study the 
differential effects of air pollution on different socio-economic status (SES) 
classes of population, which in turn raises a question of – why there might 
be inequity in air pollution? Inequity in the health effects of air pollution in 
an urban community could exist due to differences in exposure, 
susceptibility and coping capacity of different socioeconomic classes. These 
differences could be due to poor living conditions, material deprivation, 
already meager health position, heritable predilection and apprehension 
(psychosocial). Moreover, due to lower levels of social, financial and 
infrastructural amenities, individuals/ households/ communities with lower 
SES could also be confronted with lower adaptive or defensive limits to 
combat the consequences of air pollution. 

1.2. Research Evidence 

Most studies quantify inequity of health hazards from air pollution by 
examining the linkage between exposure to air pollution and socio-
economic traits, and mostly these are in the context of developed countries. 
Some other studies, on the other hand, examine SES as a modifier in the 
association between air pollution exposure and its health impact. A study 
done by Freeman (1972), using geographic coverage of three cities in the 
U.S., established that there was an inverse relation between income and SO2 
particulates exposure and the highest exposure was faced by black 
subpopulation. Gianessi et al. (1977) evaluated the distributional aspect of 
air pollution on a nation-wide basis using an integrated CB (cost benefit) 
analysis.  According to the study, standardization creates redistribution and 
the net benefits accrue to the low-income groups leading to race and 
environmental equity. A CB assessment of the impact of the uniform 
regulation by income groups confirmed a greater welfare support to blacks 
in the urban polluted residential areas. Harrison and Rubinfield (1978) tried 
to analyze the dissemination advantages from an air pollution controller 
technique for Boston using a housing value and health damages technique 
to estimate the willingness to pay benefits for 7 different income groups. 



Ecology, Economy and Society–the INSEE Journal [96] 

 

Results suggested that tangible benefits from upgrading the air quality of 
urban areas go to the poor rather than the rich, and when measured in 
pecuniary terms, including workplace benefits, it showed that distribution 
was much less towards the poor. Another evidence was given by Asch and 
Joseph (1978), who investigated both inter- and intra-city variations in air 
quality in the U.S states. Their study shows that exposure to particulates is 
associated with cities characterized by low income and hence, low 
education, low property value, and crowded population.  

Using generalized linear models (GLMs), Jerrett et al. (2004) showed that air 
pollution was related to expansive deaths in intra-urban zones of low socio-
economic characteristics. Also, low education and high manufacturing 
employment in the zones substantially enhanced mortality impacts of air 
pollution exposure. Another study by Pratt et al. (2015) studied the 
inequities in exposure to air pollution from traffic and the related risk in the 
state of Minnesota, America. They found that the risks and exposures were 
differentially larger than expected value for ethnic minorities and low SES 
population. In similar studies in Europe, Fecht et al. (2015) examined 
inequities of exposure to air pollution in England and the Netherlands at 
country, city and regional levels. Results showed a greater concentration of 
air pollution in those areas of the two countries with more than 20 per cent 
non-white population and the most deprived neighborhoods in England. A 
related study in Africa by Rooney et al. (2012) examined the spatial patterns 
of PM and its sources in four neighborhoods of varying SES in Accra using 
mixed-effects regression model and found that community SES was 
inversely associated with both PM2.5 and PM10 levels. 

In one of the Asian studies highlighting air-pollution inequity, Fan, Lam 
and Yu (2012) tried to exemplify the spatial variations in Hong Kong’s 
urban population by analyzing the relationship between SES and exposure 
to vehicular pollution. The study concluded that there was more inequality 
in private housing lands than their public counterparts. Also, older and low 
SES population faced relatively greater exposure to air pollution in contrast 
to higher SES people. However, with all residents clubbed, results showed 
no status prejudice in air pollution exposure attributable to the housing 
mechanism in Hong Kong, where the poor live in government-provided 
housing with relatively better air quality. 

There are not many studies on the inequity aspect of air pollution in India. 
Most studies primarily look at the health consequences and analysis of cost 
and benefits of improved health from a reduction in air pollution. In a 
study on Delhi, Cropper et al. (1997) estimated a dose-response function of 
health status of households to pollution levels. Results showed that more 



[97] Shivani Gupta, Sukanya Das and M.N. Murty 

 

than two per cent of non-traumatic deaths in Delhi were related to 
increased pollution levels (total suspended PM), and that this relationship 
was significant for children and adults. Similar studies were done by other 
authors. Kumar and Rao (2001) analyzed the economic benefits of 
improved air quality in Haryana and found that a representative household 
has willingness to pay (WTP) Rs. 12 to Rs. 53 per month for reducing 
particulate matter to the level of WHO standards.  

Using health production model and household survey data from Delhi and 
Kolkata, a study by Murty et al. (2003) found the annual health benefits 
from reducing air pollution levels to safe levels in the urban areas of Delhi 
and Kolkata as Rs. 4896.6 million and Rs. 2999.7 million, respectively. On 
similar lines, using a household health production model, Gupta (2006) 
examined the economic gains from reduction in air pollution in Kanpur and 
found that a typical city resident would annually save Rs. 165, if pollution 
was reduced to standard levels and the whole population of Kanpur would 
gain Rs. 213 million annually. In one of the few studies of its kind, Garg 
(2011) analyzed the equity aspects of air pollution reduction by examining 
who bears the cost of pollution abatement and who benefits from it. The 
study showed that the health effects of air pollution are more detrimental 
for the poor. The study had quantified mortality and morbidity due to 
pollution for various socioeconomic groups in Delhi with spatial data on 
concentrations of PM10 and SES. It inferred that health benefits from better 
air quality laid differentially, more with poor. A contextual study by Makri 
and Stilianakis (2008) demonstrates the vulnerability to air pollution and its 
health consequences through risk assessment.  According to this study, sub-
populace characteristics have a few socio-economic components that 
enhance vulnerability, impact on exposure, susceptibility and coping 
capacity factors. 

Another study by Kathuria and Khan (2007) investigated the relation 
between air pollution exposure and socio-economic characteristics. Using a 
two-step methodology for computing a household-specific exposure index 
for 347 houses in proximity to pollution monitoring station in Delhi, the 
study examined the relationship of air pollution exposure with socio-
economic and demographic characteristics using a multivariate regression. 
The results reflected that economically weaker sections were more exposed 
to air pollution than their counterparts, however, for socially (caste-based) 
weaker sections, the relationship with air pollution exposure was not 
significant, and no relation existed with the aspect of religion. A recent 
study by Foster and Kumar (2011) in Delhi found that people who 
remained outdoors for long hours had stronger health impacts of air 
pollution. The study also pointed out betterment in their health following 



Ecology, Economy and Society–the INSEE Journal [98] 

 

the strict regulation of air quality policies. A meta-analysis of global 
literature on air pollution by Hajat, Hsia and Neill (2015) provided overall 
similar results as presented above.  

As seen, most of the studies on India have analyzed the health impact of 
exposure to air pollution and health benefits from its reduction.  However, 
only few limited studies are available on-air pollution and its distributional 
consequences from a vulnerability context. This paper, aiming to analyze 
the distributional effects of urban air pollution in Delhi, can be a useful and 
timely study in the Indian context. It estimates a household Vulnerability 
Index for determining the role of exposure, susceptibility and coping 
capacity to air pollution. It also quantifies the relationship between 
exposure to air pollution and vulnerability in the background of socio-
economic and demographic features of the households in Delhi. Finally, it 
looks at both efficiency as well as equity aspect of reduction in air pollution 
by estimating the economic health benefits for reduction in air pollution 
and distribution of these benefits amongst different socio-economic classes 
of households in Delhi. 

 

2. MATERIALS AND METHOD 

2.1. Health Production Function Model: Theoretical Framework 

This study uses a simplified version of the health production function 
model framework for estimating vulnerability from exposure to air 
pollution and economic benefits from reduction in air pollution to safe 
limits in Delhi. As proposed by Freeman (1993), the health production 
function (HPF) and the mitigating demand function (MDF) are implicitly 
defined by an individual’s behavior who wants to maximize utility (U). On 
similar lines, as used by studies of Freeman (1993), Dasgupta (2001), Murty 
et al. (2003), Gupta (2006), Chowdhury and Imran (2010) and Adhikari 
(2012), an estimate of marginal willingness to pay (MWTP) of a 
representative urban household for reducing air pollution to safe-limits can 
be obtained as follows:1  





+


+== 1

dQ
dH

H
u

Q
M.Pm

dQ
dH.wMWTP

dQ
dI  (1) 

where U: U (Q, H, I), household utility function, health-status (H): sick-
days due to air pollution related diseases, Q: air quality, M: demand function 
for mitigation, w: wage-rate and Pm: demand price of mitigation.  

 
1 For derivation, look at Freeman (1993) 



[99] Shivani Gupta, Sukanya Das and M.N. Murty 

 

MWTP of a household for health benefits from reduction in air pollution 

)
dQ

dI(  is obtained  by the summation of three terms in equation (1): 

marginal earning that were lost due to sickness )
dQ

dH.w( , marginal 

mitigating cost )
Q

M.Pm(


  and  disutility of sickness in money terms 

).
dQ

dH
H

u(





 1 . Estimation of MWTP using this equation requires the 

estimation of simultaneous equation model consisting of health production 
function and the demand function for mitigation.  

Another possible way is to estimate a reduced form dose-response-function 
with health as a function of the physical and socio-economic aspects pooled 
with estimation of demand function for mitigation as a function of 
common set of variables as from health production function estimation 
(Freeman 1993; Gupta 2006). This requires estimating the following two 
reduced form equations of the model2: 

)Z,I,Q,Pm,w(HH =  (2) 

and 

)Z,I,Q,Pm,w(MM =  (3) 

where, Z represents a vector of other household characteristics. 

Ignoring disutility that air pollution causes, the paper estimates MWTP or 
cost of illness (COI) of a household for a small change in air pollution 
levels as a lower conservation bound:2 

Q

M
.Pm

dQ

dH
wCOI




+=  

(4) 

Usually, the dependent variable in dose-response-function is workdays lost 
but due to data unavailability number of days of sickness is taken as a 
proxy.3 

 

 
2 Here, the lower bound is due to absence of measure of loss due to averting expenditure. In 
a general health production framework, demand for averting activities is estimated 
separately. However, due to difficulty of measuring aversion behavior in monetary terms, it 
is only implicitly captured in the health production function estimation. The conservative 
bound is due to absence of disutility term which is not captured because of its difficult to 
measure tangibly. 
3 The proxy may overestimate the COI as work-days lost are also included in wage loss 
reduction )

dQ
dHw(   term in equation (4). 



Ecology, Economy and Society–the INSEE Journal [100] 

 

2.2. The Vulnerability Framework for Air Pollution 

2.2.1. Different Views on Vulnerability 

John et al. (2008) suggests that there are “three components of vulnerability, 
namely: differential exposure, susceptibility and coping mechanisms, and 
these are used to derive an initial vulnerability-framework.” IPCC (2012) 
report views vulnerability as a “universal methodology that defines 
exposure, susceptibility and societal response capacities as its factors”. 
Commission for Environmental Cooperation (2014) argues that 
“differential vulnerability to chemical exposures is characterized by the 
degree of exposure, an individual’s susceptibility to the harmful effects 
caused by the chemical and the capacity to cope with and mitigate chemical 
risks”. 

Vulnerability as an idea can be depicted by various indicators and through 
different channels. Kasperson (2002) considers four main variables which 
affect the physical as well as social characteristics of the population and 
make them “vulnerable”. These variables are (a) “susceptibility”, which 
refers to an improved probability of maintaining a longer run defensive 
impact by a sub-populace in comparison to general population; (b) 
“differential exposure” which includes the behavioral, educational, 
background and occupational exposures, along with restlessness/disutility 
which are in some cases disregarded in an evaluation; (c) “differential 
preparedness” and lastly (d) “coping ability” which includes assets of the 
communities/populace needed to survive and recover from the 
environmental impact. 

Considering all these views, Vulnerability is defined as a function of 
Exposure, Susceptibility and Coping Capacity in this paper. 

Vulnerability=f (Exposure, Susceptibility, Coping Capacity) (5) 
 

2.2.2. Vulnerability-Framework (A cobweb of factors and sub-factors)  

It’s important to note that society’s vulnerability to air pollution is intricate 
because of its multi-dimensional connections with ecological behavior 
inherent as extraneous susceptibility, differential exposure, adjusting and 
coping capacity (EPA 2003; Kasperson et al. 1995). Hence, it is useful to 
draw a vulnerability-framework template appropriate for reduced-form 
analysis, yet comprehensive to a universal problem (Turner et al. 2003). 
Many institutions and expert groups have made remarkable advancement in 
the scheming of these frameworks (Kasperson et al. 2001). 



[101] Shivani Gupta, Sukanya Das and M.N. Murty 

 

Figure 2: Vulnerability Framework: Differential Exposure, Differential 
Susceptibility and Differential Coping Capacity to Air Pollution 

Source: Adapted from John et al. (2008, 6) 

Due to the complex multifaceted relationship between factors of 
vulnerability, it becomes difficult to account for all factors and many sub-
factors, implied in the literature, in assessment of vulnerability to air 
pollution. Therefore, in order to fill this gap, this paper tries to account for 
all the three vulnerability factors (though all sub-factor may not be 
considered due to data and other constraints). In line with Kasperson et al. 
(1995), EPA (2003), Turner et al. (2003), and John et al. (2008), this paper 
derives the required modified and simplified version of vulnerability-
framework for air pollution using the factors of health production model 
framework explained above, using the framework in Figure 2. 

As seen in Figure 2, there are numerous channels of assessing vulnerability 
to air pollution. Regarding susceptibility, the population characteristics 
which could lead to differential health impacts of air pollution and other 
chemical encounters are mainly – underlying poor health (past resistant 
reactions) and disease state, conceivably hereditary predilection, race and 
gender distribution. This commonly cited variable, “susceptibility” can be 
internal – if it originates from factors such as age, gender, birth, health 
defects or race (which is the root for genital sensitivities), as well as external 



Ecology, Economy and Society–the INSEE Journal [102] 

 

– which is mainly due to health status, bad habits (such as drinking, 
smoking, etc.) that could lead to chronic diseases, obesity, etc. (Morata et al. 
1997; EPA 2003; John et al. 2008; Lipfert 2004; Stilianakis 2015). 
Concerning differential exposure to air pollution, it could be due to  
behavioral exposures from unhealthy living conditions, exposure to busy 
traffic/rail road and long hours spent outdoor, indoor air pollution 
exposure caused by poor fuel type and/or appliances, occupational 
exposures due to informal or unorganized sector jobs, and finally, exposure 
due to educational status, family background and deprivation of assets 
(EPA 2003; John et al. 2008; Lipfert 2004; Stilianakis 2015). Finally, the 
coping capacity, which is the ability to incur abatement expenditures, could 
be dependents on sub-factors such as socio-economic position, health care 
access (like medical insurance), awareness of air-borne diseases, economic 
status (through employment status and educational level) and governmental 
support (Deguen and Zmirou-Navier 2010; EPA 2003; John et al. 2008 and 
Stilianakis 2015). 

Establishing a vulnerability definition along with its parameters remains 
essential for its assessment. Indicators and sub-indicators have been 
established based on existing literature, available data, expert judgments and 
modeling capabilities of the health production function model. This 
framework will further be used to construct a household specific 
vulnerability index, which is a feature scale score calculation of sub-
indicators (as independent variables of the health production and mitigating 
demand function models), indicators (as estimated dependent variables of 
the health production and mitigating demand function models) and final 
vulnerability index score. The methodology for the construction of this 
index is given in section 2.4. 

2.3. Sample Design and Method of Data Collection 

Information on average PM10 concentration was collected from Central 
Pollution Control Board (CPCB), Delhi for 3 months (October-December 
2015). As compared to average concentrations of NO2 and SO2, which were 
below safe levels, PM10 level was much above the safe limits (100 µg/m3) of 
National Ambient Air Quality Standards (NAAQS) and was therefore 
mainly considered for this study. Of the 17 Pollution Monitoring Stations 
(PMS) across Delhi, 11 stations provided consistent levels of PM10 data.  
This study is based on PM10 data from 10 PMS that capture diverse 
locations in the West, East, South and North Delhi, both residential as-
well-as industrial areas including: Anand Vihar, Mandir Marg, R.K. Puram 
and Punjabi Bagh monitored by Delhi Pollution Control Board in addition 



[103] Shivani Gupta, Sukanya Das and M.N. Murty 

 

to I.T.O., Siri Fort, Pitampura, Shahzada Bagh, Janakpuri and Shahadra 
monitored by CPCB.  

Data on health-status, socio-economic and demographic factors was 
collected through a household survey conducted in January 2016. A pilot 
field test was conducted in December 2015 to refine as well as fine-tune the 
draft questionnaire. For a comparative analysis of the inequity aspect of air 
pollution, the survey sample households were identified from slums, middle 
income group areas and high-income group areas. For this, a two-stage 
stratification was followed as done by Gupta (2006) and Adhikari (2012). In 
Stage-1, an equal number of households were identified within 2 km radius 
of each of the 10 PMS. In Stage-2, an equal number of households under 
different categories were identified on the basis of their locality and 
accommodation type: high class (independent housing), service class (flat 
type housing) and slums across each PMS. In this study, 180 households 
were surveyed; 18 households within 2 km radius of each of 10 PMS. 
Further, each set of these 18 households were classified into low, middle 
and high socio-economic categories based on the location and 
accommodation setting (6 random households for each category under each 
PMS). Data on health-status, mitigating expenditure as well as averting 
activities to combat air- pollution impact was collected in the recall period 
of three months, also capturing information on household members 
suffering from any chronic disease and their awareness about air pollution 
related diseases. Comprehensive recall data included – number of days of 
sickness for each member, household medical expenses (medical expenses, 
including medicines, doctor’s fee, cost of diagnostics, travel, etc., of air 
pollution related ailments), family health insurance and treatment type.  
Pollution aversion activities included family opting to stay indoors, 
travelling extra kilometers, using gas masks, air purifier, air conditioner (AC) 
transport and other activities such as using a scarf or closing door/windows 
to avoid air pollution. 

For housing characteristics, information collected included – household 
type, ownership, drinking water, number of rooms and so on. Data on 
utilization of air conditioner, type of cooking fuel, heaters was also collected 
to analyze indoor air-quality. For socio-economic characteristics, 
information obtained included – caste, religion, family background, educa-
tional attainment, hours spent outside, unhealthy habits (smoking, drinking, 
lack of exercise, etc.), occupational profile, wage/income and job being 
indoor/outdoor. For demographic characteristics, details included – age 
and gender, household size, number of working and school going family 
members and so on. Lastly, households’ monthly income and other 



Ecology, Economy and Society–the INSEE Journal [104] 

 

indicators of wealth, i.e., durable assets owned, and monthly family 
expenditure were also collected. 

Another survey of 10 random doctors was conducted in February 2016 to 
elicit information on severity of health impact (on the population in 
general) of the 11 chronic diseases that were taken in household survey 
using a Likert Scale (1-3). 

2.4. Econometric Model Specification: Household Health Production 
Function Model  

In the health production function model used in this paper, number of sick 
days in the household and medical expenses due to air pollution are 
considered as dependent variables. The data collected for family sick days 
represent count data and data on medical expenses observed zero censor 
for several data points. As in studies by Gupta (2006), Chowdhury and 
Imran (2010), and Adhikari (2012), Poisson Regression (PR) model4 and 
Tobit-Regression model could be used for estimation of health status and 
mitigating expenditure, respectively. However, the assumption of equality 
of conditional mean and variance acts too restrictive for PR models and is 
unable to account for over-dispersion of data, leading to a huge standard 
error of estimation which was corrected using Negative Binomial 
Regression (NBR) model. Owing to an unequal mean (53.66) and variance 
in health status suggesting over-dispersion in data, NBR was used for 
estimating health production function over PR and was validated by a 
likelihood-ratio test. Also, the data showed only few occurrences with zero 
mitigating-expenditure (only 16 per cent of sample) and a near normal-
distribution with log-transformation. Hence, a log form Ordinary Least 
Square Regression (OLSR) Model was selected for modelling the mitigating 
demand function over Tobit-regression. Empirically, this study estimates 
the household HPF and MDF as represented by the two reduced form 
equations with health -status, and mitigating expenditure, both dependent 
on a common set of independent variables: 

Reduced Form Equation 1: HPF  

H = h0 + h1logHEIi + h2CIi + h3HABi + h4FAMi + h5SESi + 
h6SUFFERi + h7AWAREi + h8AVERTi + h9SOCi + 
h10INSi + h11GENDERi + ui 

(6) 

 
4 The Poisson Model is given by: (Prob Hi = hi/xi) = , such that Hi=0, 1, 2.... is 

a count variable that captures numbers of days of sickness in ith household and hi is the 

mean as-well-as variance of sick days. 



[105] Shivani Gupta, Sukanya Das and M.N. Murty 

 

Reduced Form Equation 2: MDF 

LogM = m0 + m1logHEIi + m2CIi + m3HABi + m4FAMi + 
m5SESi + m6SUFFERi + m7AWAREi + m8AVERTi + 
m9SOCi + m10INSi + m11GENDERi + ei 

(7) 

where ui and ei represent error terms  

These two equations are estimated using cross-sectional regressions - NBR 
used to estimate the health status and log-form OLSR to estimate the log 
(Mitigating Expenditure). The description of variables used are as follows: 

2.4.1. Dependent Variables 

Health-Status(H): The sum of number of days of sickness due to air 
pollution related diseases in each household over the recall period of three 
months.  

Mitigating-Expenditure(M): The medical expenses of the household in the 
recall period of three months due to air pollution related diseases on 
medicines, doctor fees, journey cost to doctor and diagnostic tests. 

2.4.2. Independent Variables 

Household Exposure Index (HEI)5: It was constructed using data on mean 
PM10 (µg/m3) concentration of the location-specific PMS, household 
working profile and hours spent outside home by working members 
assuming differential exposure to air pollution for working members (11 
hours) followed by school/college going members (9 hours) and non-
working (2 hours). The average HEI was 117.84 µg/m3 against PM10 
average of 360.38 µg/m3

. Amongst the PMS’, Anand Vihar, on an average, 
was the most polluted whereas Mandir Marg was the least polluted in the 
study period (see Figure 3). 

Chronic Diseases Index (CI):  Weighted average of the prevalence of 11 
chronic diseases in a particular house including - bones problem, 
pneumonia, eye-disease, thyroid, diabetes, bronchitis bp, tb, asthma, cancer 
and heart-disease. Weights are based on the severity score of the diseases 
calculated using the doctor survey. 

 
5The HEI is calculated for particular station and household as:  
HEIij = Pi*(Nw*Hw+Nnw*Hnw+NSOC*HSOC/24*Family Size) in ith household and jth 
pollution monitoring station, where P: Mean PM10 (µg/m3) Nw: working members, NSOC: 
school/college going members, Nnw: non-working members whereas Hw, HSOC and Hnw 
represent the average hours spent outside by the working, school/college going members, 
respectively. 



Ecology, Economy and Society–the INSEE Journal [106] 

 

Figure 3: Average concentrations of PM10 and Household Exposure Index in 
µg/m3 for 10 pollution monitoring stations in Delhi (October to December 
2015) 

 

Source: CPCB, Delhi and Authors’ Calculation 

Habits (HAB): Unhealthy habits of a household as the ratio of number of 
family members with unhealthy habits scaled to family size. These habits 
could either be smoking or drinking or not exercising for any member.  

Socio-Economic Status Index (SES): One of the most important variables 
for the study was SES index which represented the household’s social and 
economic status, computed using a Principal Component Analysis (PCA)6 
method. It facilitated easy interpretation as it presented the data under a 
single-dimension rather than multi-dimensional framework (Cordova 2009). 

The variables used for construction of the SES Index, were captured using 
highly correlated variables under three household characteristics namely: 
wealth, housing and socio-economic characteristics (Table 1). The data was 
transformed by dividing it into three quantiles based on the SES Index 
where top, medium and bottom 33.33 percent were generated to 
differentiate between the high, middle and lower SES households, 
respectively. 

 
6 In PCA, weights are assigned to each observed variable summed up by determining the 
data variance direction, as the set of correlated variables are transformed into an 
uncorrelated component set. PCA captures the maximum amount of information that is 
identical for all variables to be used as an index that is linear for all variables and provides 
appropriate weights for every variable such that the index captures the largest variations.  



[107] Shivani Gupta, Sukanya Das and M.N. Murty 

 

Table 1: Socio-Economic Status Index Variables 

Wealth Characteristics 

TV  Washing Machine  Micro-Oven  

Fridge  A.C  Heater  

Geyser  Music System Car  

Laptop Phones 
 

Housing Characteristics 

Rooms  Water Facility  

Socio-Economic Characteristics 

Education  Better Off (Location)  Organized Sector  

Source: Authors' Own compilation 

Avert (AVERT): The following were taken as a score measured from a 
household’s strategies for averting air pollution: staying indoors, travelling 
extra kilometers, using gas masks, air purifier, ac-cars v/s public-transport 
and other activities such as using a scarf or closing door/windows of the 
households to avoid air pollution.   

Gender (GENDER): The ratio of females to family-size. 

School/College Going (SOC):  The ratio of school or college going 
members to family-size. This is used in order to capture a demographic 
factor in the modelling. 

Awareness Index (AWARE): It is defined as respondent’s awareness of the 
number of diseases out of the 17 diseases known to be clinically caused by 
air pollution. 

Suffer (SUFFER): It is the ratio of members suffering from air pollution 
related diseases to family-size. It is used as a physical health control variable. 

Health Insurance (HI): A dummy variable defined as ‘1’ if the household 
has family health insurance, and ‘0’ otherwise. 

Family Size (FAM): Number of members in a household. 

2.4.3. Vulnerability Index for Air Pollution 

The vulnerability framework (explained above in Figure 2) is used to build a 
household-specific vulnerability index (V) in order to examine the role of 
exposure, susceptibility and coping mechanism in determining vulnerability 
to air pollution. This comprises 2 components: model-estimated values of 

HPF represented by Ĥ and negative of the estimated value of MDF 

represented by 
^

LM . Here, vulnerability is dependent directly on 



Ecology, Economy and Society–the INSEE Journal [108] 

 

susceptibility factor, captured by Ĥ  and negatively dependent on coping-

capacity factor, captured by
^

LM . The exposure factor in case of vulnerability 
to air pollution is not easy to estimate alone, as it does not act as a 
behavioral variable directly (also due to unavailability of weather-related 
data) but has linkage to susceptibility and coping capacity factor of air 
pollution. Hence, the exposure factor of vulnerability to air pollution is 
captured implicitly as household exposure index by health production and 
mitigating demand functions, both. The vulnerability index is calculated as a 
function of the health production function, which captures the exposure 
and susceptibility factors, and the mitigating-expenditure which will capture 
the exposure and coping-capacity factors of vulnerability. 

Unfortunately, no available study on air pollution computes a vulnerability 
index at a household level.  However, many studies, such as by Christenson 
et al. (2016), Kissi at al. (2015), Baeck (2014) and Kallis et al. (2010) have 
computed vulnerability index for climate change impacts (heat stroke, flood 
risks, etc.) at a regional or city level. In line with these studies, this study has 
developed a household-specific vulnerability index/score7 for air pollution 
using the following formula: 

Vulnerability Score = Susceptibility Score + (-) Coping Capacity 
Score 

(8) 

Feature scaling score method is used for standardizing the factor scores of 
susceptibility and coping-capacity8 making them comparable to get an 
accurate value of vulnerability index as is done by UNDP (2002) in building 

the Human Development Index. This is done since Ĥ  has a count unit 

whereas 
^

LM  has rupees unit. 

The feature scaling is done using the formula given below:  

Feature Scale Score = 
)Value(Min)Value(Max

)Value(MinValue

−

−
 

(9) 

 
7 Vulnerability Index is calculated with the supposition that each of the factors have an equal 
role in determining vulnerability with only contrast being that exposure factor is being 
captured implicitly in Susceptibility and Coping-Capacity scores. 
8 Susceptibility score and (-) coping capacity score are given by following equations: 

Susceptibility score =
)Ĥ(Min)Ĥ(Max

)Ĥ(MinĤ

−

−  and (-) Coping Capacity score = 

)LM(Min)LM(Max

)LM(MinLM
^^

^^

−

− . 

Here, the Ĥ represents the predicted value of health status whereas 
^

LM  represents negative 
of the predicted value of Log (Mitigating Expenditure). 



[109] Shivani Gupta, Sukanya Das and M.N. Murty 

 

The summary statistics shows that the mean Vulnerability Index is 0.76. 

2.5. Empirical Model of Inequity in Air Pollution  

As discussed through literature, air pollution inequity captured by 
vulnerability is principally dependent on SES of the subpopulation. For a 
statistical analysis, the paper uses a multivariate regression model with 
dependent as vulnerability index on independents as socio-economic 
characteristics (SEC), demographic characteristics (DEC) along with other 
control variables: 

Vij = f (SECij, DECij, Controlsij) (10) 

where i=1, 2….180 Households and j=1, 2….10 PMS.  

For this study, the empirical specification of the model is given by: 

V = v0 + v1logHE1 + v2SESi + v3GENDERi + v4HIi + 
v5INFOUTDOORi + v6BGi + v7ADULTSi + zi 

(11) 

where, zi is the error term. Socio-economic characteristics of the household 
are captured by the SES Index and Family Background (BG) as a dummy, 1 
for urban and 2 for rural. Demographic characteristics are captured by age 
(ADULTS) as the proportion of members who are 15 years or older and 
gender (GENDER). The Household Exposure Index (HEI) measuring the 
exposure factor was only implicitly accounted in calculating the vulnerability 
index and hence is taken as an independent control here. Other control 
variables taken into consideration are informal outdoor occupation 
(INFOUTDOOR) given by the ratio of members with informal outdoor 
occupations (like petty traders, construction workers, etc.) to working 
members and lastly, a Health Insurance (HI) dummy.  

The descriptive statistics of the dependent and independent variables used 
in all the estimation models are provided in Table A1 in Appendix. 

 

 

 

 

 

 

 

 



Ecology, Economy and Society–the INSEE Journal [110] 

 

Table 2: Estimates of the Health Status (H: Family Number of Days in 
Sickness) 

Variable  Negative Binomial Coefficients  

Health Status: Dependent Variable  

Log (Household Exposure Index)  0.45**= h1 

Chronic Index  0.16**  

Socio-Economic Status Index  -0.083*** 

Habits  0.39***  

Family Size  0.087**  

Suffer  1.16***  

Awareness Index  -0.039  

Avert Activities  -0.138***  

School/College Going  0.87***  

1. Health Insurance  0.016  

Gender  0.093  

Constant 0.46  

Ln alpha -0.76*** 

Observations 180 

Source: Authors' Own Calculation 
Legend: * p<0.10; ** p<0.05; *** p<0.01. Likelihood-ratio test of alpha=0: 
chibar2 (01) = 3194.63 Prob>=chibar2 = 0.000 

3. RESULTS AND DISCUSSIONS 

3.1. Household Health Production Function Model 

Results of the estimated HPF using the NBR9 model are given in Table 2. It 
shows that eight out of eleven parametric estimates have significant relation 
to health status with expected directions. The coefficient of variable HEI is 
positive while the coefficient of variable SES Index is negative as expected 
and they are significant at 5 per-cent and 1 per-cent levels, respectively. 

Table 3 provides the results of the estimated mitigating expenditure 
function. This reduced form function has common independent variables 
with HDF function. Ten of eleven independent variables in the model have 
coefficients with expected signs and significant at 1 to 5 per cent levels.  
The independent variables HEI and SES Index have positive coefficients as 
expected which are significant at 5 per cent and 1 per cent levels, 
respectively. The awareness index has a negative coefficient which is 
significant at 5 per cent level. 

 
9 The significant p value of LR-test (as given in Table 2) suggests that NBR is better over a 
PR Model validating the results of estimation. 



[111] Shivani Gupta, Sukanya Das and M.N. Murty 

 

Table 3: Estimates of the Log Mitigating Expenditure: Log (M) 

Variable  OLS (Log-Form) Coefficients  

Log (Mitigating Expenditure): Dependent Variable  

Log (Household Exposure Index)  0.44**= m1 

Chronic Index  0.089  

Socio-Economic Status Index  0.16***  

Habits  0.39***  

Family Size  0.16***  

Suffer  1.23***  

Awareness Index  -0.103** 

Avert Activities  -0.29*** 

School/College Going  1.03***  

1.Health Insurance  0.32**  

Gender  0.75**  

Constant 4.92***  

Observations 177 

Source: Authors' Own Calculation 
Legend: * p<0.10; ** p<0.05; *** p<0.01 

Figure 4: Scatter Plot between SES Quantiles and Vulnerability Index 

 

         Source: Authors' Own Calculation 

3.2. Estimates of Empirical Model of Inequity in Air Pollution 

Estimates show that lower SES households have a mean vulnerability score 
of 0.97 while the higher SES households have a mean vulnerability score of 
only 0.58. Figure 4 shows that the households in the lower SES quantile 
group have higher vulnerability scores as compared to households in the 
middle and higher SES quantile group. 



Ecology, Economy and Society–the INSEE Journal [112] 

 

Table 4: Estimates of the Vulnerability Index (V) 

Variable OLS Coefficients 

Vulnerability Index: Dependent Variable 

Log (Household Exposure Index) 0.061*** 

Socio-Economic Status Index -0.032*** 

Gender -0.18*** 

Health Insurance -0.094*** 

Informal Outdoor Occupation 0.065*** 

Background 0.041* 

Age 0.0302 

Constant 0.55*** 

Observations 180 

Source: Authors' Own Calculation 
Legend: * p<0.10; ** p<0.05; *** p<0.01 

Table 4 provides the estimates of vulnerability index function. Results show 
that as socio-economic status of a household decreases, vulnerability to air 
pollution increases significantly (1 per-cent significance-level). As expected, 
the parameters of independent variables - pollution exposure, proportion of 
members with informal outdoor occupation, absence of health insurance 
and rural family background - are positively and significantly associated with 
vulnerability to air pollution. However, age shows no significant effect on 
vulnerability. The parameter gender reflects an unexpected result: it shows 
that with increased proportion of men in family, vulnerability to air 
pollution increases which might be because men have a higher working 
exposure than women in Delhi. The results as hypothesized, found 
presence of air pollution inequity in the context of Delhi. 

3.3. Estimates of Monetary Benefits from Reducing Air-Pollution to 
Safe Limits in Delhi 

Using equation (4) of the household health production model explained 
above, the cost of illness (COI) or household health cost of air pollution 
could be estimated as: 

HEI

M

HEI

H
.WCOI




+




=  

(12) 

In other words, the health benefits of reduced air pollution consist of 
reduced cost of work days lost in the family and the reduced mitigating 
expenditure. 



[113] Shivani Gupta, Sukanya Das and M.N. Murty 

 

The welfare gains to a typical household from reduced pollution are 
estimated considering sample mean values of variables: household exposure 
index (HEIM), mitigating expenditure (MM), family size (FAMM)), health 
status (HM), family daily wage rate (WM), and household exposure index safe 
(HEI.SAFEM)10. Table A2 in Appendix provides descriptive statistics 
variables used in the estimation. 

The methodology for calculating total annual benefits to a typical 
household is given as follows: 

A typical household’s marginal willingness to pay for reduction in air 
pollution exposure by 1µg/m3 for last three months recall is calculated by 
summing up the marginal effect of reduction in wage loss11 and the 
mitigating expenditure in last 3 months (recall period) quarter given by: 














+














=

M

M

M

M

M

M

HEI

M
m

HEI

H
h

FAM

W
QMWTP 11  

(13) 

where h1 and m1 represent the estimated coefficient of household exposure 
index for health production and mitigating demand function regression 
model, respectively. 

Now, a typical household’s marginal willingness to pay for reduction in air 
pollution exposure by 1µg/m3 is annualized by weighting12 the last quarter 
(3 months recall period) as compared to the other three quarters of the year: 

 
10 The household exposure index at safe level of PM10 is given by  
HEI.SAFEij = 100*(Nw*Hw+Nnw*Hnw+NSOC*HSOC/24*Family Size) in ith household and jth 

PMS , where 100 µg/m3 is the safe PM10 limits given by CPCB (NAAQS), Nw: working 
members, NSOC: school/college going members, Nnw: non-working members whereas Hw, 
HSOC and Hnw represent the average hours spent outside by the working, school/college 
going members, respectively. 
11 The marginal effect of reducing air pollution exposure (from reduced work days lost) by 1 
unit for a typical household is multiplied by family wage rate and divided by family size in 
order to get the benefits for just the working members. This is because health status has 
been taken as proxy of work days lost. 
12 For annualizing the marginal willingness to pay, a rank sum method was incorporated to 
calculate the weight for the three months recall period data. Since, the household survey was 
conducted around winter season (October-December 2015), studies and air pollution 
seasonal trend analysis by CPCB has shown that winter season traditionally has been the 
most severe in terms of air quality levels. According to the report “NAQI Status of Indian 
Cities in 2015-16” by CPCB, the national air quality of Delhi around winter season was seen 
in severe/ worst category. Hence, the highest rank was assumed for the study period last 
quarter out of the four yearly season quarters and was used to calculate weight for it. The 



Ecology, Economy and Society–the INSEE Journal [114] 

 

40.

QMWTP
MWTP =   

(14) 

For calculating the total annual economic benefits for a typical household 
from reducing air pollution exposure from current (study period) level 
(HEI) to safe level (HEI.SAFE at PM10 of 100µg/m³), the MWTP for 1 
unit fall in air pollution exposure is multiplied by the change in exposure 
from current level to safe level: 

Total Annual Economic Benefit = MWTP * (HEIM – HEI.SAFEM) (15) 

For inferring economic benefits for the whole population in Delhi, 
extrapolation is done using data of Census (2011) in India, which shows the 
population in Delhi as 0.0168 billion and workforce participation rate 
(percentage of people employed in the population) as 33.4 per cent. The 
particulars of the calculated economic benefits from reduction in air 
pollution are provided in Table A2 in Appendix. 

The above formulation and information gives the annual economic benefits 
from the reduction in workdays (sick days taken as a proxy here 13) lost 
from pollution exposure to safe levels (100 µg/m3) as Rs. 28,820.35 for a 
typical-household, and as Rs. 6,080.24 for a representative working 
individual.  Entire working population of Delhi gets benefits worth Rs. 
34.11 billion from reduced workdays lost, if a typical household’s benefit is 
extrapolated using working participation rate as 33.4 percent (Census 2011). 
Similarly, the annual economic benefits from reduced mitigating expenses 
for a typical household due to reduction in air pollution exposure to safe 
level is estimated as Rs. 5,157.77 and Rs. 1,088.13 for a representative 
individual. The extrapolated benefits for entire Delhi population from 
reduced mitigating expenses are estimated as Rs. 18.28 billion. Hence, the 
total annual monetary gain from reduced air pollution exposure (from 
current to safe level) to the National Capital of India is Rs. 52.4 billion. To a 
typical household and an individual, the total annual gain amounts to Rs. 
33,978.12 and Rs. 7,168.37, respectively. An estimate of share of annual 
benefits from reduced air pollution to safe levels comes to about 2.54 
percent of annual income of a typical Delhi household. 

 

 
formulae used for weight of the last quarter is (Rank of the last Quarter/Sum of all 

ranks) 40
10

4
.==  

13 The proxy may overestimate the economic benefits, as work-days lost are also included in 

wage loss reduction ).(
dQ

dHW term shown in equation (12). 



[115] Shivani Gupta, Sukanya Das and M.N. Murty 

 

3.4. Distribution of Economic Benefits 

The annual economic benefits to a typical household from reduced air 
pollution exposure from current to safe levels (100 µg/m3) are quite high as 
shown above. However, the question that remains is - who actually gains 
the most from pollution reduction? To answer this question, the shares of 
total annual economic benefits for the three different SES quantile of the 
households (Low, Middle and High SES) in Delhi are estimated. Figure 5 
provides the distribution of economic benefits due to reduction in air 
pollution amongst different socio-economic classes of households in Delhi. 

As shown in Table 5, estimates of annual economic benefits from reduction 
in pollution exposure to safe levels for a typically lower socio-economic 
household is Rs. 8,701.33 which is much lower than that of a higher SES 
household which saves as high as Rs. 50,631.95, while a typical middle-class 
household gains a mediocre amount of Rs. 28,376.92. This means that in 
absolute terms a higher SES household gains the most out of pollution 
exposure reduction. This estimate though essential does not show the true 
and real depiction of the distribution of gains and reality could be inferred 
by looking at the gains in relative terms, estimating the annual benefits share 
as a percentage of annual income for the three classes of households. These 
estimates show that a lower SES household saves 4.96 percent out of their 
annual income which is much larger than a higher SES household saving 
which is merely 1.97 per cent out of the annual family income, if pollution 
exposure is reduced to safe levels. For a middle-class family the saving rate 
out of income is estimated as 2.23 per cent. 

Table 5: Distribution of Economic Benefits by Reduction in Air-Pollution 
amongst different Socio-Economic Classes of Households 

SES Annual Benefit Annual Benefit Share (%) 

Low Rs. 8,701.33 4.96 

Middle Rs. 28,376.92 2.23 

High Rs. 50,631.95 1.97 

Source: Authors’ own Calculation 

 

4. CONCLUSIONS AND POLICY SUGGESTIONS 

The econometric results show that ceteris paribus, households with lower 
SES have a higher vulnerability to air pollution and its consequences. 
Vulnerability is more prominent for household members with informal 
outdoor occupations and families who have higher air pollution exposure. 



Ecology, Economy and Society–the INSEE Journal [116] 

 

Moreover, households with health insurance have lower vulnerability to air 
pollution which could be due to lower mitigating expenditure and better 
coverage of health facilities. However, age shows no significant effect on 
vulnerability. These results conclude that households/communities with 
lower educational attainment, pronounced deprivation of wealth assets, 
higher proportion of people in unorganized job sectors and poor living 
conditions (such as living near busy roads or in slums, having unhealthy 
water facilities) will be the most vulnerable to air pollution and its impacts. 
The results are in line with the findings of Landrigan and Goldman (2011) 
and Hajat et al. (2015). 

Estimates of economic benefits (in terms of health) from air pollution 
reduction to safe limit (100µg/m3) from current levels (study period) for a 
typical household amount to Rs. 33,978.12 and for the entire population of 
Delhi it is Rs. 52.4 billion. A typical household can save about 2.54 per cent 
per year of their annual income from reduction in pollution exposure to 
safe levels.   

It can be concluded that reduction in pollution (in Delhi) can have critical 
monetary benefits through improved health status and lowered medical 
expenses on air pollution related illness by the households and the 
population of Delhi. There are already a few other studies done in the past 
which show similar results, though, with much lower annual benefits in the 
context of other cities. For instance, Gupta (2006) found the economic 
benefits for the city of Kanpur to be around Rs. 200 million per year. This 
highlights the plausibility of a high estimated value of economic benefits for 
the whole population, given Delhi being amongst the most polluted as well 
as highly populated cities in the world. However, these estimates only 
provide information about the gains to a typical household or population 
from pollution reduction. The critical question is who gains the most from 
economic benefits of pollution reduction and whether the policies that lead 
to air pollution reduction can also lead to “pro-equity” benefits to all socio-
economic segments along with overall “pro-efficiency” benefits to a typical 
household of Delhi. Distribution of the economic gains from air pollution 
reduction reflects that in absolute terms a typically higher SES gets higher 
annual benefits which may be since these have higher annual income and 
hence will save more in absolute terms. In reality, looking at the gains in 
relative terms by estimating the annual benefits share as a percentage of 
annual income for the three classes of households, estimates shows that a 
lower SES household saves the maximum proportion of its annual income, 
i.e., 7.93 per cent as compared to higher SES household for which annual 
benefit share is merely 3.15 per cent of the annual family income if 
exposure to air pollution is reduced to safe limits.  



[117] Shivani Gupta, Sukanya Das and M.N. Murty 

 

Thus, it can be concluded that a lower SES household with a higher 
vulnerability to air pollution will benefit the most from reduced air 
pollution (in relative terms). Studying the impact of socio-economic 
position on vulnerability to air pollution can help assess policy implications 
in tackling the menace of air pollution and population’s vulnerability to it. A 
cooperative strategic policy can help economically poor households use 
appropriate defensive behavior and activities to reduce the severity of air 
pollution consequences and thus, reduce inequity in the environmental 
aspect along with the reduction in air pollution. 

In recent years, several significant studies have addressed policy suggestions 
for air pollution in the context of Delhi. A recent source apportionment 
study by TERI (2018) identified multiple sources for the variation in PM2.5 
and PM10 concentrations during summer and winter months - transport, 
biomass burning, industries and dust being the significant ones. It proposed 
several interventions, ranging from liquified petroleum gas penetration, 
introduction of gaseous fuels and enforcement of new and stringent 
SO2/NOx/PM2.5 standards for industries using solid fuels, inspection and 
maintenance of vehicles, strong implementation of BS-VI (Bharat Stage) 
norms to congestion management, etc. Further, attention has been given at 
the national and sub-national level to issues of proper management of crop-
residue burning such as implementing blanket ban, promoting subsidies and 
compensation programs for farmers to buy advanced and modern 
equipment for in situ crop residue management, initiatives to diversify the 
crop production and straw management, creating market for agri-residues, 
while focus given to embedded socio-economic, cultural and behavioral 
elements responsible for widespread practice of burning of agri-residues has 
been limited (Kumar et al. 2015).  

Looking at the course of action from the distributional aspect, social 
variables that have prompted air pollution inequity as shown in the study 
should additionally be given due importance along with those policies that 
will help in reduction of air pollution. Ujjwala scheme has been promoted 
by the government for higher penetration of liquified petroleum gas in 
the households from lower SES. More attention is needed for reducing 
occupational exposures, thereby, improving the working conditions of 
people. Legal structures also need to be effectively designed to protect 
workers in outdoor jobs and other vulnerable groups. Another strategy 
prospect could be more focused awareness creation about consequences of 
air pollution and procurement of more information on the health impacts 
of air pollution, which affect individual’s adaptive and mitigating capacity to 
combat with it.  



Ecology, Economy and Society–the INSEE Journal [118] 

 

This study found it important to analyze efficiency as well as equity matters 
of air pollution in Delhi. A general observation made by authors such as 
Blank (2002) is that the trade-off between equity and efficiency only exists 
in the shorter run (as efficiency can only be persistent in the longer run 
when these are balanced); in the long run, both go hand-in-hand (Berg and 
Ostry 2011; Krugman 2014). In terms of air pollution inequity, the gap 
between the two can be minimized by implementing pro-poor policies with 
support from the community and government, through policies that 
mandate efficient pollution control standards and finally, through 
redistribution of payments via- subsidies, health insurance facilities and 
increasing education investment to reduce both distributional vulnerability 
to air pollution along with reduction in overall exposure. 

Vulnerability to air pollution and its consequences cuts across various 
sectors at different scales, and thus calls for coherence and integration at 
the policy level to deal with such an imminent problem. This study serves as 
a potential for reviewing new developments in further research and use the 
outcomes of the analysis for policy implications that act as a toolkit for the 
government at a micro level to base decisions at a macro level in the longer 
run. 

 

REFERENCES 

Adhikari, N. 2012. Measuring the health benefits from reducing air pollution in Kathmandu 
Valley. Kathmandu: South Asian Network for Development and Environmental 
Economics. 

Asch, P., and J. S. Joseph. 1978. “Some Evidence on the Distribution of Air 
Quality.” Land Economics 54: 279-297. https://doi.org/10.2307/3145999 

Baeck, S. H., S. J. Choi, G. W. Choi, and D. R. Lee. 2014. “A study of evaluating 
and forecasting watersheds using the flood vulnerability assessment index in 
Korea.” Geomatics, Natural Hazards and Risk 5 (3): 208-231. 
https://doi.org/10.1080/19475705.2013.803268 

Berg, A., and J. Ostry. 2011. “Inequality and Unsustainable Growth: Two Sides of 
the Same Coin? ”, IMF Staff Discussion Note, SDN/11/08. 
https://doi.org/10.5089/9781463926564.006 

Blank, R. M. 2002. “Can equity and efficiency complement each other?” Labour 
Economics 9 (4): 451-468. https://doi.org/10.1016/S0927-5371(02)00011-8 

Cardona, O.D., M.K. van Aalst, J. Birkmann, M. Fordham, G. McGregor, R. Perez, 
R.S. Pulwarty, et al. 2012. “Determinants of risk: Exposure and vulnerability. 
Managing the Risks of Extreme Events and Disasters to Advance Climate Change 
Adaptation.” In A Special Report of Working Groups I and II of the Intergovernmental Panel 

https://doi.org/10.2307/3145999
https://doi.org/10.1080/19475705.2013.803268
https://doi.org/10.5089/9781463926564.006
https://doi.org/10.1016/S0927-5371(02)00011-8


[119] Shivani Gupta, Sukanya Das and M.N. Murty 

 

on Climate Change (IPCC), edited by C.B. Field, V. Barros, T.F. Stocker, D. Qin, D.J. 
Dokken, K.L. Ebi, M.D. Mastrandrea, et al. 65-108. Cambridge, UK, and New 
York, USA: Cambridge University Press. 

Central Pollution Control Board. 2000. “Air Quality Status and Trends in India.” 
National Ambient Air Quality Monitoring Series. Delhi: Ministry of Environment 
and Forests. 

Central Pollution Control Board. 2001. “Toxic Air Pollutants.” Parivesh Newsletter. 
Delhi: Ministry of Environment and Forests. 

Central Pollution Control Board. 2015-16. “NAQI Status of Indian Cities in 2015-
16”. Accessed on 26th October 2018. 
http://cpcb.nic.in/cpcbold/upload/Latest/Latest_119_NAQI-StatusIndianCities-
2015-16.pdf  

Chowdhary, T and M. Imran. 2010. “Morbidity Costs of Vehicular Air Pollution: 
Examining Dhaka City in Bangladesh”. Kathmandu: South Asian Network for 
Development and Environmental Economics. 

Christenson, M., S. D. Geiger, J. Phillips, B. Anderson, G. Losurdo, and H. A. 
Anderson. 2016. “Heat Vulnerability Index Mapping for Milwaukee and 
Wisconsin”.  Journal of public health management and practice 23 (4):396-403. 
https://doi.org/10.1097/PHH.0000000000000352 

Commission for Environment Cooperation. 2014. “A Framework Document: 
Factors to Consider in Characterizing Vulnerability to Environmental 
Contamination Across North America”. Montreal, QC:55. This report was written 
by Mario Rivero Huguet for the Secretariat. 

Cordova, A. 2009. “Methodological Note: Measuring Relative Wealth Using 
Household Asset Indicators”. Insights Series No. I0806. 
https://www.vanderbilt.edu/lapop/insights/I0806en_v2.pdf 

Cropper, M., N.B. Simon, A. Alberini, S. Arora, and PK Sharma. 1997. “The 
Health Benefits of Air Pollution Control in Delhi”. American Journal of Economics 79 
(5): 1625-1629. https://doi.org/10.2307/1244393 

Dasgupta, P. 2001. “Valuing Health Damages from Water Pollution in Urban Delhi: A 
Health Production Function Approach”. Working Paper. New Delhi: Institute of 
Economic Growth. 
http://www.iegindia.org/upload/publication/Workpap/wp210.pdf 

Deems, H. J. 2010. “Vulnerability of rural communities in the Mediterranean region 
to climate change and water scarcity: The case of Cyprus”. Masters' Thesis. 
Universitat Autónoma de Barcelona. 
https://www.cyi.ac.cy/images/projects/eewrc/clico/Deems_Rural_Vulnerability_
Cyprus_thesis_Sep2010.pdf  

http://cpcb.nic.in/cpcbold/upload/Latest/Latest_119_NAQI-StatusIndianCities-2015-16.pdf
http://cpcb.nic.in/cpcbold/upload/Latest/Latest_119_NAQI-StatusIndianCities-2015-16.pdf
https://doi.org/10.1097/PHH.0000000000000352
https://doi.org/10.2307/1244393
https://www.cyi.ac.cy/images/projects/eewrc/clico/Deems_Rural_Vulnerability_Cyprus_thesis_Sep2010.pdf
https://www.cyi.ac.cy/images/projects/eewrc/clico/Deems_Rural_Vulnerability_Cyprus_thesis_Sep2010.pdf


Ecology, Economy and Society–the INSEE Journal [120] 

 

Deguen, S., and D. Zmirou-Navier 2010. “Social Inequalities Resulting from 
Health Risks Related To Ambient Air Quality: A European Review”. The European 
Journal of Public Health 20(1): 27-35. https://doi.org/10.1093/eurpub/ckp220 

Fan, X., K. C. Lam, and Q.Yu. 2012. “Differential exposure of the urban 
population to vehicular air pollution in Hong Kong”. Science of the Total Environment, 
426: 211-219. https://doi.org/10.1016/j.scitotenv.2012.03.057 

Fecht, D., P. Fischer, L. Fortunato, G. Hoek, K. De Hoogh, M. Marra, and A. 
Hansell. 2015. “Associations Between Air Pollution and Socioeconomic 
Characteristics, Ethnicity and Age Profile of Neighborhoods in England And the 
Netherlands”. Environmental Pollution 198: 201-210. 
https://doi.org/10.1016/j.envpol.2014.12.014 

Foster, A., and N. Kumar. 2011. “Health effects of air quality regulations in Delhi, 
India”. Atmospheric Environment 45(9) :1675-1683. 
https://doi.org/10.1016/j.atmosenv.2011.01.005 

Freeman, A. M. 1972. “Distribution of Environmental Quality”. In: Environmental 
Quality Analysis edited by A. Kneese and B. Bower. Baltimore: John Hopkins Press.  

Freeman, A. M. III. 1993. The Measurement of Environmental and Resource Values: Theory 
and Methods. Washington, D. C: Resources for the Future. 

Garg, A. 2011. “Pro-Equity Effects of Ancillary Benefits of Climate Change 
Policies: A Case Study of Human Health Impacts of Outdoor Air Pollution in New 
Delhi”. World Development 39(6): 1002-1025. 
https://doi.org/10.1016/j.worlddev.2010.01.003 

Gianessi, L. P., H.M. Peskin, and E. Wolff. 1977. “The Distributional Implications 
Of National Air Pollution Damage Estimates”. In The Distribution of Economic Well-
being edited by F.T. Juster. 85-102. Cambridge: Ballinger Publishing.  

Gupta, U. 2006. “Valuation of Urban Air Pollution: A Case Study of Kanpur City 
in India”. Kathmandu: South Asian Network for Development and Environmental 
Economics. 

Hajat, A., C. Hsia, and M.S. O’Neill. 2015. “Socioeconomic Disparities and Air 
Pollution Exposure: A Global Review”. Current Environmental Health Reports 2:440-
50. https://doi.org/10.1007/s40572-015-0069-5 

Harrison, D. and D. L. Rubinfeld .1978. “The Distribution of Benefits from 
Improvements in Urban Air Quality”. Journal of Environmental Economics and 
Management 5: 313-332. https://doi.org/10.1016/0095-0696(78)90017-7 

https://www.who.int/phe/health_topics/outdoorair/databases/AAP_database_re
sults_2014.pdf?ua=1  

Jerrett, M., R. T. Burnett, J. Brook, P. Kanaroglou, C. Giovis, N. Finkelstein, and B. 
Hutchison. 2004. “Do Socioeconomic Characteristics Modify the Short-Term 
Association Between Air Pollution and Mortality? Evidence from a Zonal Time 

https://doi.org/10.1093/eurpub/ckp220
https://doi.org/10.1016/j.scitotenv.2012.03.057
https://doi.org/10.1016/j.envpol.2014.12.014
https://doi.org/10.1016/j.atmosenv.2011.01.005
https://doi.org/10.1016/j.worlddev.2010.01.003
https://doi.org/10.1007/s40572-015-0069-5
https://doi.org/10.1016/0095-0696(78)90017-7
https://www.who.int/phe/health_topics/outdoorair/databases/AAP_database_results_2014.pdf?ua=1
https://www.who.int/phe/health_topics/outdoorair/databases/AAP_database_results_2014.pdf?ua=1


[121] Shivani Gupta, Sukanya Das and M.N. Murty 

 

Series in Hamilton, Canada”.  Journal of Epidemiology and Community Health 58: 31-40. 
https://doi.org/10.1136/jech.58.1.31 

John, J., M. Matooane, R. Oosthuizen and C.Y. Wright. 2008. “Vulnerability to air 
pollution: To intervene or not to intervene”. 
https://www.researchgate.net/publication/30510866_Vulnerability_to_air_polluti
on_To_intervene_or_not_to_intervene  

Kasperson, J. X. and R.E. Kasperson. 2001. SEI Risk and Vulnerability Programme 
Report 2001-01. Stockholm: Stockholm Environment Institute. 

Kasperson, J. X., R. E. Kasperson, and B. L. Turner, II, eds. 1995. Regions at Risk: 
Comparisons of Threatened Environments. Tokyo: United Nations University Press. 

Kasperson, R. E. 2002. Personal communication to Michael Callahan, Chair of the 
Cumulative Risk Technical Panel, as part of an EPA Science Advisory Board 
Consultation on the draft Framework for Cumulative Risk Assessment. 

Kathuria, V., and N. A. Khan. 2007. “Vulnerability to Air Pollution: Is There Any 
Inequity in Exposure?”.  Economic and Political Weekly 42 (30): 3158-3165. 

Kissi, A. E., G. A. Abbey, K. Agboka, and A. Egbendewe. 2015. “Quantitative 
Assessment of Vulnerability to Flood Hazards in Downstream Area of Mono Basin 
South-Eastern Togo: Yoto District”. Journal of Geographic Information System 7(6): 607-
19. https://doi.org/10.4236/jgis.2015.76049 

Krugman, P. 2014. “Inequality Is a Drag.” New York Times, August 7, 2014. 
http://www.nytimes.com/2014/08/08/opinion/paul-krugman-inequality-is-a-
drag.html?_r=0 

Kumar, P., S. Kumar, and L. Joshi. 2015. Socioeconomic and Environmental Implications 
of Agricultural Residue Burning. New Delhi: Springer. https://doi.org/10.1007/978-
81-322-2014-5 

Kumar, S and D.N. Rao. 2001. “Valuing Benefits of Air Pollution Abatement 
Using Health Production Function: A Case study of Panipat Thermal Power 
Station, India”. Journal of Environmental & Resource Economics 20: 91-102. 
https://doi.org/10.1023/A:1012635627808 

Landrigan, P. J., and L. R. Goldman, 2011. “Children’s vulnerability to toxic 
chemicals: a challenge and opportunity to strengthen health and environmental 
policy”. Health Affairs 30(5): 842-850. https://doi.org/10.1377/hlthaff.2011.0151 

Lipfert, F. W. 2004. “Air pollution and poverty: does the sword cut both ways?” 
Journal of epidemiology and community health 58(1): 2-3. 
https://doi.org/10.1136/jech.58.1.2 

Morata, T. C., A. C. Fiorini, F. M. Fischer, S. Colacioppo, K. M. Wallingford, E. F. 
Krieg, and C. L. G. Cesar. 1997. “Toluene-Induced Hearing Loss Among 
Rotogravure Printing Workers”. Scandinavian Journal of Work, Environment & Health 
289-298. https://doi.org/10.5271/sjweh.222 

https://doi.org/10.1136/jech.58.1.31
https://www.researchgate.net/publication/30510866_Vulnerability_to_air_pollution_To_intervene_or_not_to_intervene
https://www.researchgate.net/publication/30510866_Vulnerability_to_air_pollution_To_intervene_or_not_to_intervene
https://doi.org/10.4236/jgis.2015.76049
https://doi.org/10.1007/978-81-322-2014-5
https://doi.org/10.1007/978-81-322-2014-5
https://doi.org/10.1023/A:1012635627808
https://doi.org/10.1377/hlthaff.2011.0151
https://doi.org/10.1136/jech.58.1.2
https://doi.org/10.5271/sjweh.222


Ecology, Economy and Society–the INSEE Journal [122] 

 

Murty, M. N., S. C. Gulati and A. Banerjee. 2003. “Health Benefits from Urban Air 
Pollution Abatement in the Indian Subcontinent.” Discussion Paper no. 62/2003. 
Delhi: Institute of Economic Growth.  

Pratt, G. C., M. L. Vadali, D. L. Kvale, and K. M. Ellickson. 2015. “Traffic, Air 
Pollution, Minority and Socio-Economic Status: Addressing Inequities in Exposure 
and Risk”. International Journal of Environmental Research and Public Health 12(5): 5355-
5372. https://doi.org/10.3390/ijerph120505355 

Rooney, M. S., R. E. Arku, K. L. Dionisio, C. Paciorek, A. B. Friedman, H. 
Carmichael, and M. Ezzati. 2012. “Spatial and Temporal Patterns of Particulate 
Matter Sources and Pollution in Four Communities in Accra, Ghana”. Science of the 
Total Environment 435: 107-114. https://doi.org/10.1016/j.scitotenv.2012.06.077 

Stilianakis, N. I. 2015. Susceptibility and Vulnerability to Health Effects of Air Pollution: 
The Case of Nitrogen Dioxide. Brussels: European Union. 

TERI. 2018. “Source Apportionment of PM2.5 & PM10 of Delhi NCR for 
Identification of Major Sources”. http://www.teriin.org/sites/default/files/2018-
08/AQM-SA_0.pdf  

Turner, B. L., R. E. Kasperson, P. A. Matson, J.J. McCarthy, R.W. Corell, L. 
Christensen, and C. Polsky. 2003. “A Framework for Vulnerability Analysis in 
Sustainability Science”. Proceedings of the National Academy of Sciences 100(14): 8074-
8079. https://doi.org/10.1073/pnas.1231335100 

U.S. EPA. 2003. Framework for Cumulative Risk Assessment. EPA/600/P-02/001F. 
Washington, DC: U.S. Environmental Protection Agency, Office of Research and 
Development, National Center for Environmental Assessment. 

World Bank. 2015. Leveraging Urbanization in South Asia: Managing Spatial 
Transformation for Prosperity and Livability. Washington, DC: World Bank.  

World Health Organization. 2014. “WHO’s Ambient Air Pollution database ‐ 
Update 2014".  

https://doi.org/10.3390/ijerph120505355
https://doi.org/10.1016/j.scitotenv.2012.06.077
http://www.teriin.org/sites/default/files/2018-08/AQM-SA_0.pdf
http://www.teriin.org/sites/default/files/2018-08/AQM-SA_0.pdf
https://doi.org/10.1073/pnas.1231335100


[123] Shivani Gupta, Sukanya Das and M.N. Murty 

 

Appendix  

Table A1: Descriptive Statistics of Variables used in Estimation of Health 
Production Function and Mitigating Demand Function Model, and in Estimation 
of Vulnerability to Air Pollution Model 

Variable Mean Std. Dev. Min Max 

Health Status  53.66  50.58  0 255 

Mitigating Expenditure  6506.63  9317.5 0 60950 

Vulnerability Index 0.76 0.197 0.35 1.37 

Household Exposure Index  117.054  44.96  46.15  288.304 

Chronic Index  0.531  0.69  0 4.3 

Socio-Economic Index  -9.31e-10 3.206  -4.48  11.39  

Habits  0.74  0.54  0 3 

Family Size  4.74  1.77  1 11 

Suffer  0.63  0.29  0 1 

Awareness  10.47  2.52  0 16 

Avert  1.73  1.36  0 5 

School/College Going 0.24  0.2006  0 0.75  

Health Insurance  0.55  0.49  0 1 

Gender  0.49  0.18  0 1 

Outdoor Job  0.29  0.41  0 1 

Background  1.32  0.47  1 2 

Adults  0.81  0.19  0.33  1  

 

 

 

 

 

 



 

Table A2: Estimates of Total Annual Economic Benefits and Annual Benefit Share (%) for Reduction in Air-Pollution 

h1 m1 Mean Values 

Last Three Months FAMᴹ Wᴹ INCᴹ 

D
el

h
i 
P

o
p

u
la

ti
o

n


 

(B
ill

io
n

) 

D
el

h
i 
W

o
rk

in
g 

P
ar

ti
ci

p
at

io
n

 R
at

e 
* 

E
st

im
at

ed
 c

o
ef

fi
ci

en
t 

o
f 

H
o

u
se

h
o

ld
 E

x
p

o
su

re
 I

n
d

ex
 f

o
r 

H
P

F
 r

eg
re

ss
io

n
 m

o
d

el
 

E
st

im
at

ed
 c

o
ef

fi
ci

en
t 

o
f 

H
o

u
se

h
o

ld
 E

x
p

o
su

re
 I

n
d

ex
 f

o
r 

M
D

F
 r

eg
re

ss
io

n
 m

o
d

el
 

HEIᴹ 
HEI. 

SAFEᴹ 
Hᴹ Mᴹ 

F
am

ily
 S

iz
e 

F
am

ily
 W

ag
e 

R
at

e 
(R

s.
) 

A
n

n
u
al

 H
o

u
se

h
o

ld
 I

n
co

m
e 

(R
s.

) 

H
o

u
se

h
o

ld
 

E
xp

o
su

re
 I

n
d

ex
 

H
o

u
se

h
o

ld
 

E
xp

o
su

re
 I

n
d

ex
 

S
af

e 

F
am

ily
 D

ay
s 

o
f 

S
ic

k
n

es
s 

M
it

ig
at

in
g 

E
xp

en
d

it
u
re

 (
R

s.
) 

0.45 0.44 117.05 32.7 53.67 6506.63 4.74 3139.64 1337399 0.0168 33.4 

 

Annualized Economic Benefits (Rs.) to a typical household for reduction in pollution exposure by 1 µg/m3 

Gain from Reduction in Sick Days  
(For Working Family Members) 

Gain from Reduction in Mitigating 
Expenses (For all Family Members) 

Annual Willingness to Pay  
(Household) 

Rs. 341.67 Rs. 61.14 Rs. 402.82 

 

Annualized Economic Benefits (Rs.)  to a typical household for reduction in pollution exposure from current to safe level 
(100 µg/m3, PM10) 



[125] Shivani Gupta, Sukanya Das and M.N. Murty 

 

Gain from Reduction in Sick Days  
(For Working Family Members) 

Gain from Reduction in Mitigating 
Expenses (For all Family Members) 

Total Annual Economic Benefits 
(Household) 

Rs. 28820.35 Rs. 5157.77 Rs. 33978.12 

 

Annualized Benefits Share (%) to a typical household for reduction in pollution exposure from current to safe 
level (100 µg/m3, PM10) 

4.06% 

 

Annualized Economic Benefits (in Rs. Billion) to Delhi Population for reduction in pollution exposure from current to 
safe level (100 µg/m3, PM10) 

Gain from Reduction in Sick Days  
(For Working Population) 

Gain from Reduction in Mitigating 
Expenses (For Entire Population) 

Total Annual Economic Benefits  
(Delhi) 

Rs. 34.11 Billion Rs. 18.28 Billion Rs. 52.4 Billion 

Source: Authors’ Calculation 

 Population and Working Participation Rate, Census (2011) 

 

 

 


