Energy Policy 140 (2020) 111397

Contents lists available at ScienceDirect

Energy Policy
journal homepage: http://www.elsevier.com/locate/enpol

Towards a new approach in measuring energy poverty: Household level

analysis of urban India☆
Hippu Salk Kristle Nathan a, *, Lakshmikanth Hari b
a
Institute of Rural Management Anand (IRMA), Anand, India
b
Centre for Sustainable Development, K J Somaiya Institute of Management, Somaiya Vidyavihar University, Mumbai, India

A R T I C L E I N F O A B S T R A C T

JEL classification: Measurement of energy poverty has been a missing priority in energy research. There is hardly any consensus
Q4 regarding whether to consider energy as a resource or capacity, output or outcome, and quantity or service. In
I32 this paper, we revisit major approaches in literature to assess energy poverty: economic, engineering, and access-
Keywords: based, including others. We critique the economic and engineering approaches for the arbitrariness of cut-offs
Energy access and the misplaced emphasis on less meaningful energy quantity. Finding the access-based approach most suit­
Measuring energy deprivation
able, we propose a method of assessing energy poverty based on deprivation in modern cooking and lighting
Energy poor and its sub-categories
Depth and severity of energy poverty
fuels. We divide the energy-poor into three groups: transitional, moderate, and extreme energy-poor. We apply
Complementary poverty measures the method to India’s urban areas of different states and union territories using the national sample survey data.
The results reveal energy poverty to be primarily dictated by deprivation in cooking and a greater incidence of
energy poverty in larger states. In contrast to conventional measures, we compute the depth and severity of
energy poverty only for energy-poor as opposed to for the entire population. These complementary measures
depict energy poverty of energy-poor more accurately and therefore will direct the attention of pro-poor energy
policies and programmes appropriately.

1. Introduction is no single internationally-accepted definition of modern energy, In­

Energy poverty can be defined as ‘the absence of sufficient choice in
accessing adequate, affordable, reliable, high-quality, safe, and envi­
ronmentally benign energy services to support economic and human
development’ (Reddy, 1994). It refers to a situation where the house­
holds have lack access to modern energy services and that has adverse
impacts on quality of life and overall development and equity in a so­
ciety (UNDP, 2009; Maxim et al., 2016). While acknowledging that there
ternational Energy Agency indicates households with modern energy
have minimum level of electricity, safe as well as sustainable cooking
and heating fuels along with efficient stoves, energy-enabled productive
activities (e.g. mechanical power for agriculture, home-based enter­
prise, etc.), and public services (street lights, hospitals, schools, etc.)
(IEA, 2018).
The importance of a study on energy poverty may be attributed to
primarily four reasons. First and foremost, unlike income poverty, there

☆
The authors have benefited from comments by three anonymous reviewers and presentations of the earlier versions of the paper at the 8th INSEE Conference at
the Indian Institute of Science, Bangalore; seminar talks at Acharya Bangalore Business School, Bangalore; Asia Development Bank Institute, Tokyo; Center of
Development Studies, University of Cambridge, Cambridge; College of Engineering and Technology, Bhubaneswar; CUTS International, Jaipur; India Observatory,
London School of Economics and Political Science (LSE), London; Gujarat Institute of Development Research, Ahmedabad; Indian Institute of Management,
Ahmedabad; Nabakrushna Choudhury Centre for Development Studies (NCDS), Bhubaneswar; National Institute of Advanced Studies (NIAS), Bangalore; and PVP
Siddhartha Institute of Technology, Vijayawada; and training course lecture at NIAS-Department of Science and Technology Training Programme, NIAS, Bangalore.
Earlier versions of the paper have been published as working papers at NCDS, Bhubaneswar and Institute of Rural Management, Anand. A part of this work was
carried out when the first author was at India Observatory at LSE, London, UK as a Subir Chowdhury Fellow on Quality and Economics from February to May 2017,
working on the research topic “Energy Poverty and Quality of Life in India - A household-level analysis”. Discussion with Dr. Ruth Kattumury, financial support from
the Fellowship, and administrative support by Mr. Kevin Shields during the first author’s stay at LSE are gratefully acknowledged. We are thankful to Ms. Indrani
Talukdar for editorial suggestions.
* Corresponding author.
E-mail addresses: hippu@irma.ac.in, happyhippu@gmail.com (H.S.K. Nathan), hari@somaiya.edu, hlkanth@yahoo.com (L. Hari).

https://doi.org/10.1016/j.enpol.2020.111397
Received 19 June 2019; Received in revised form 24 February 2020; Accepted 28 February 2020
Available online 20 March 2020
0301-4215/© 2020 Elsevier Ltd. All rights reserved.
H.S.K. Nathan and L. Hari
is no consensus on how to measure energy poverty (Barnes et al., 2014;
Halff et al., 2014).1 This calls for a review of merits and demerits of
different metrics of energy poverty and finding ways of reconciling their
rationalities through a new measure. The second important reason for
the study of energy poverty stems from the magnitude and seriousness of
the problem (Sovacool, 2014). Depending on the definition, one-third to
one-fourth of humanity is energy-poor (Halff et al., 2014). The third
reason pertains to the acknowledgement that measuring energy poverty
is the essential first step in tackling the problem. Birol (2014) wrote
quite appropriately: ‘in life what gets measured gets managed’. The
fourth reason is more fundamental involving the reality that access to
energy is not simply an outcome of development but also a means to it.
Energy acts as an engine for the production of goods and services and is
vital for social progress including provision of health care, water, sani­
tation, and education (ADB, 2007). Besides, it has a multiplier effect on
the productivity of income-generating activities in agriculture, industry,
and services (Modi et al., 2005; ADB , 2007; IAEA , 2008; Birol, 2014;
Nathan and Raj, 2016). Access to energy that is affordable, sustainable,
and reliable is crucial to poverty eradication, socio-economic trans­
formation, and overall sustainable development (UN, 2013; Common­
wealth Secretariat, 2014). Given the countless ways in which energy
enhances the lives of the poor and enables development (UN, 2005,
2010),2 provisioning of modern energy services has become a ‘moral’
imperative (Birol, 2014) and, in this sense, a poverty metric based on
energy deprivation as opposed to income or wealth is more relevant for
policy makers and implementers of development policies (Halff et al.,
2014).
Regardless of its importance, energy poverty has remained a missing
global priority until recently. Highlighting its lack of attention on the
part of experts and policymakers, Birol (2007, 2014) attributes energy
poverty to the poverty of energy economists; and its eradication is first a
statistical battle.3 Barring some discrete individual efforts from the likes
of Prof. A.K.N. Reddy of India and Prof. J. Goldemberg of Brazil, uni­
versal energy access has not been on the global agenda. It could not
figure among the Millennium Development Goals (Halff et al., 2014). It
was only in 2002 that the International Energy Agency (IEA)’s World
Energy Outlook Report assessed energy poverty for the first time. This
missing priority has received some recent amendments: In 2012, the
United Nations celebrated the International Year for Sustainable Energy
for All (UN, 2011). Next, the United Nations General Assembly declared
2014–2024 as the decade of sustainable energy for all (UN, 2012).
Subsequently, energy was included as a goal among Sustainable
Development Goals (SDGs); Goal 7 reads: Ensure access to affordable,
reliable, sustainable, and modern energy for all (SE4All, 2014; UNDP,
2015).
The delay in conceptualizing and measuring energy poverty mainly
involves consensus deficit regarding situating energy in the course of
development. Whether one should consider energy as a ‘resource’ or
‘capacity’, ‘output’ or ‘outcome’, and ‘quantity’ or ‘service’, the
conundrum is perennial. Energy in the conventional or physical sense is
considered a fuel or input resource whereas under Sen’s capability
approach nomenclature it may be considered as an enabler or capacity ‒
both from the perspective of households and society at large (Halff et al.,
1
Income poverty is ‘usually based on measure of minimum consumption of
food and non-food items necessary to sustain life’ (Halff et al., 2014).
2
Lighting extends the study hours of children and work hours for adults;
modern cooking saves one from exposure to indoor air pollution; refrigeration
can preserve food and medicine; energy can improve access to water, mobility,
and information and communication, enhance quality of products, and create
employment opportunities and increase wages (Birol, 2014).
3
By statistical battle, the author indicates the lack of agreement on mea­
surement of energy access.
2
Energy Policy 140 (2020) 111397
2014).4 To measure energy requirements, one can measure the quantity
of energy demand. However, a more useful approach would be to
measure energy services.5 For instance, for lighting, energy quantity is
measured in terms of electricity consumed, whereas energy service is
measured in terms of luminance, i.e., the service desired by the user
from electricity. The same quantity of energy may yield different mag­
nitudes of services depending on the efficiency of the device. For
example, using the same quantity of electricity, an incandescent bulb
will give poorer lighting services compared to a compact fluorescent
lamp. Energy may also be measured in terms of output (for instance,
access to electricity) or outcome (welfare gains because of the electricity
access that likely to result in an increase in study hours for children and
working hours for adults) (Barnes et al., 2014).
Against this backdrop, the importance of conceptualization and
measurement of energy poverty cannot be over-emphasized. This paper
evaluates the major approaches in literature of assessing energy poverty,
viz., economic, engineering, and access-based approaches, including
others (Pachauri and Spreng, 2004; Pachauri et al., 2004; Kemmler,
2007; Culver, 2017; Jain and Shahidi, 2019). It argues that the economic
approach, which is based on income poverty, suffers from the arbitrar­
iness of poverty line that divides the people discretely as poor and
nonpoor; hence this approach may not be very reasonable. The engi­
neering approach is a bottom-up and a more detailed way of assessing
energy requirement based on household needs. However, this approach
suffers from the limitation of data availability, which is expected to be
extremely dynamic. Also, there is a mismatch between energy quantity
(that the engineering approach generally focusses on) and energy ser­
vice (that’s the concern of households) because of different efficiencies
of energy devices for the same service. Moreover, like the economic
approach, the engineering approach is not free from the arbitrariness of
cut-off.
This paper proposes an access-based method of computing energy
poverty. It departs from convention in two ways. First, in order to reduce
the arbitrariness of poverty line cutoff and resulting discreteness, it di­
vides the energy-poor into three groups: extreme energy-poor, moderate
energy-poor, and transitional energy-poor. As a second point of depar­
ture the paper, taking a cue from a recent exercise by Nathan (2018),
computes depth and severity of energy poverty only for the energy-poor,
not the entire population. The paper presents, thereby, depth and
severity as complementary measures to the incidence of energy poverty.
Incidence gives the share of population that is energy-poor, while depth
and severity give level of poverty of energy-poor accounting for their
gap and distribution, respectively. The paper provides an empirical
illustration of the proposed methodology by considering energy poverty
in different entities (states and union territories) of India for urban areas.
The rest of the paper is organized as follows: Section 2 briefs the
major approaches of measuring energy poverty. Section 3 outlines the
new methodology proposed in this paper. Section 4 provides the
empirical illustration. Section 5 provides concluding remarks.
2. Approaches of measuring energy poverty
2.1. Economic approach
The economic approach to energy poverty is linked to economic
parameters. The most common factor of such an approach is mapping
4
For details see introduction to the capability approach in Stanford Ency­
clopedia of Philosophy (2011).
5
Energy services are the services the energy devices provide such as lighting,
heating for cooking, space heating, etc. (Modi et al., 2005; Kowsari and Zerriffi,
2011).
H.S.K. Nathan and L. Hari
energy poverty to income (or expenditure) poverty (Pachauri et al.,
2004).6 Under this approach, energy poverty line is considered as the
value of average energy consumption of households who are at income
poverty line (Foster et al., 2000; Pachauri et al., 2004).7 There are two
other alternatives under economic approach. One is based on the energy
budget share, i.e., share of household expenditure spent on energy,
where one finds poor spending greater share on energy compared to
higher income groups (Leach, 1987; Pachauri et al., 2004; Kemmler,
2007).8 The other is based on the effective price, i.e., price paid by the
household per unit useful energy,9 which shows a wicked gradient:
better-off paying a lower price per unit energy (Leach, 1987; Foster
et al., 2000; Pachauri et al., 2004; Kemmler, 2007). Lack of access to
affordable modern energy services encourages the poor to use inefficient
fuels along with inefficient devices and spend more time as well as effort
in availing the same (Pachauri et al., 2004; Kemmler, 2007). It is
worthwhile to note that in the ‘poverty line’ approach energy-poor lies
below the cut-off, whereas in the ‘budget share’ and ‘effective price’
approaches the energy-poor lies above the cut-offs.
The economic approach has some serious shortcomings. First and
foremost, the economic parameters on which the energy poverty mea­
surement is based, such as income or price, are continuous variables
while discrete cut-offs for identification of the poor is arbitrary (Nathan,
2018). Watts (1968, p.325), in the context of income poverty line, wrote
quite succinctly: “Poverty is not really a discrete condition. One does not
immediately acquire or shed the afflictions we associate with the notion
of poverty by crossing any particular income line.” Many authors have
questioned the rationale of unduly placing great value on increment in
income that moves an individual from below to above the poverty line,
while giving absolutely no value to income increments below the
poverty line (Easterly, 2009). So, assuming the poor as one homogenous
group is disputable. Similarly, categorizing households adjacent to a
cut-off on either side into two extreme groups – poor and nonpoor – is
unreasonable given the similarities of their consumption patterns
(Nathan, 2018). Consider, for instance, the Government of UK’s earlier
definition that identified a household to be energy-poor when it has to
spend more than 10% of its income on energy to heat its home to an
adequate standard (Boardman, 1991; DTI, 2006; Moore, 2012; Aguilar
et al., 2019); the categorization of two households spending a little less
and a little more than the cut-off, say 9.9% and 10.1% of their income for
the same purpose, as nonpoor and poor, respectively, is nothing short of
6
A borader alternative to this approach is to map energy poverty to
thresholds in human development index (HDI) or different human development
parameters (Goldemberg and Johansson, 1995). A different method relating to
the income proposed by Khandker et al. (2012) considers energy poverty to the
level of energy consumption (at low income scenario) that remains insensitive
to income change.
7
For instance, in Foster et al. (2000), the average energy consumption of
households falling within a 10% range of official economic poverty line have
been considered to determine energy or fuel poverty line.
8
Other than specifying simple income share cut-off on energy expenditures,
there are a few variants like ‘double median expenditure’ or 2M: the cut-off is at
the double the median share of households’ energy expenditure (Schuessler,
2014; Aguilar et al., 2019); ‘minimum income standard’ or MIS: the net income
(total income less by expenditure on energy and housing) below MIS defined by
Bradsahw et al. (2008) (Aguilar et al., 2019); ‘after fuel cost poverty’ or AFCP:
the net income lower than 60% of all households’ median net income (Hills,
2012; Aguilar et al., 2019); ‘low income high cost’ or LIHC: with two cut-off
conditions-energy expenditure higher than median and income condition
same as AFCP; and ‘compound energy poverty indicator’ or CEPI: with same
two cut-offs as LIHC with the first one with additional ‘or’ condition that energy
expenditure not lower than 10% of income (Aguilar et al., 2019).
9
Useful energy is defined as “the energy effectively made available to the
user in terms of the services delivered through end-user equipment” (Madur­
eira, 2014).
3
Energy Policy 140 (2020) 111397
arbitrary.10
Another important shortcoming of the economic approach is that the
measure may not be indicative of the actual poverty situation. In the
poverty line approach, lower energy consumption may be the result of
using higher efficient devices. Hence, as reasoned earlier, quantity of
energy consumption is less meaningful.11 Similarly, energy poverty may
not dictate a larger budget share on energy need; this could occur due to
luxurious and wasteful consumption or larger household size (Foster,
2000). A higher effective price per unit of energy may indicate that the
household’s consumption occurs at a higher range: for instance, per unit
electricity price may be greater at higher consumption ranges (Bachao,
2018). Another undesirable implication of the economic approach is
that exclusive reliance on income makes income poverty a proxy for
energy poverty making energy poverty measure mostly redundant
thereby.
2.2. Engineering approach
The engineering approach of assessing energy poverty is basically a
bottom-up approach that estimates directly the energy requirement of
households based on normative requirements of different energy ser­
vices on the part of households. It calculates the poverty threshold by
considering the specifications of different energy carriers (calorific value
of fuels) and energy appliances (size, efficiency, etc.) used to derive
these basic energy services (Pachauri et al. 2004; Goozee, 2017). For
instance, Reddy (1999) estimated that the basic final energy re­
quirements were 100 W per capita assuming safe, clean, and efficient
cooking with liquefied petroleum gas (LPG) or LPG-like fuel, and elec­
tricity for lighting, space comfort, food preservation, and entertain­
ment.12 Other examples of estimations of energy poverty based on
engineering approach are: 1 kW per capita primary energy by Gold­
emberg et al. (1985), 50 kg of oil equivalent (kgoe) of annual com­
mercial energy per capita (40 kgoe for cooking and 10 kgoe for lighting)
by Modi et al. (2005), among others.13 Another example under this
approach is the minimum annual electricity requirement specified by
IEA (2014, p.81) to be 250 kWh and 500 kWh in rural and urban
households, respectively. This approach has the flexibility of finding
energy thresholds for different needs in different geographical and
socio-cultural settings.14 For different past initiatives based on the en­
gineering method see Pachauri et al. (2004), Swan and Ugursal (2009),
and Culver (2017), among others.15
The engineering approach has several limitations. This method is
computationally intensive requiring complex and inter-related data on
household dwelling units, occupants, fuels, and appliances (Swan and
Ugursal, 2009; Goozee, 2017; Culver, 2017). In addition, the basic needs
are subjective to consumers and may vary with season, region, and
climate (Pachauri et al., 2004; Culver, 2017). Hence, these data are
10
This arbitrariness has been highlighted in Hills (2012), Aguilar et al. (2019),
.
11
Conventionally, the cut-offs are based on final energy, i.e., the energy
delivered to the input devices a consumer uses (Madureira, 2014); hence, do
not consider for the efficiencies of the devices. Useful energy accounts, on the
contrary, for device efficiencies. We acknowledge that improvement in
energy-efficiency does not lead to proportionate decrease in energy consump­
tion because of rebound effect (Gillingham et al., 2014).
12
Watts/capita is an alternative unit of expressing energy requirements. For
example, the IEA (2014) annual requirement of 250 kWh electricity for a rural
household can be expressed as: 250*1000/(365*24*5) ¼ 69 W/capita consid­
ering a month to be 30 days and assuming size of the households to be five.
13
Instead of single number, Pachauri and Spreng (2004) has proposed
different thresholds for different combinations of energy sources.
14
For these advantages, Pachauri et al. (2004) considers this approach to be
robust.
15
In Culver (2017), engineering method comes under approach of measuring
inputs in terms of energy.
H.S.K. Nathan and L. Hari
extremely dynamic and their availability always remains a tricky issue.
Also, comprehensive and expensive surveys conducted to collect such
data can become obsolete with changing technologies, preferences, and
practices. Another limitation of the engineering approach is that it fails
to account for consumer behavior and socio-economic choices con­
cerning energy consumption (Goozee, 2017). The focus of the engi­
neering approach is on quantity of energy, which is not a very useful
indicator as discussed earlier.
Besides, the engineering approach is flawed with the limitation of
cut-offs and is not, in this sense, free from arbitrariness. This arbitrari­
ness is a serious issue because, like income, quantity of energy is a
continuous variable. Hence, a discrete cut-off is not very reasonable.
Take, for instance, when you consider the minimum yearly requirement
of 500 kWh of electricity for an urban household as poverty cutoff.
Households consuming a slightly lower, say 499 units per year and
slightly higher, say 501 units per year, may reflect similar energy
characteristics but are forced into categories of poor and nonpoor, in
that order.
2.3. Access-based approach
The access-based approach of measuring energy poverty depends on
whether a household has access to desirable energy services. The IEA
uses the access-based approach to assess worldwide energy poverty by
calculating the number of people lacking access to electricity and
modern cooking fuel (IEA, 2017). This approach has a binary nature that
identifies households as with and without access to energy services
(Culver, 2017). In this sense, this method suits for the discreteness
requirement in categorizing people as poor and nonpoor.
The access-based approach is presented in certain literature, as in the
case of Pachauri et al. (2004), as a complicated method evincing diffi­
culty in data finding. This is so because access is defined in such liter­
ature as physical access that includes not only data on physical access of
energy resources, but also data on access to equipment along with the
market prices of electricity and electrical equipment, households pur­
chasing capacity, quality of supply, and so on (Pachauri and Spreng,
2004; Pachauri et al., 2004). When access is considered as mere con­
nections, non-inclusion of data on quantity of supply, quality of the
service, its affordability, etc. is treated as limitations of the approach
(Culver, 2017).
Following access-based approach, there are aggregated measures of
energy poverty at household level. ‘Multi-dimensional Energy Poverty
Index’ or MEPI proposed by Sadath and Acharya (2017) and Acharya
and Sadath (2019) is one such example which aggregates household’s
energy deprivation in cooking and lighting fuels using the AF framework
multi-dimensional poverty (Alkire and Foster, 2011). Nussbaumer et al.
(2012) is another such example which used the same MEPI term and
same framework, but extended the measure from deprivations in only
cooking and lighting to include deprivations in appliances for preserving
food (fridge), education or entertainment (radio/television), and tele­
communication services (landline/mobile phone). However, these
aggregated measures though are conceptually attractive, but less
meaningful and interpretable (Culver, 2017), because of loss of infor­
mation in the process of aggregation (Nathan et al., 2015).
2.4. Other approaches
An alternative approach to the above-mentioned three approaches
can be assessing energy poverty based on the outcomes of energy ser­
vices Culver (2017). Outcomes of energy access do not confine to the
after-impacts of energy use, and can include efforts incurred in pro­
curing the energy means. ‘Energy Poverty Index’ or EPI conceived by
Mirza and Szirmai (2010) is an example, which is an combined index of
different outcome-centric parameters like frequency of procuring energy
sources, time-spent, distance-travelled, and means of transport used in
doing so, involvement of household members in procurement, and
4
Energy Policy 140 (2020) 111397
impact on health because of use of energy sources. But, such combined
indices have the drawback of loss of information (Nathan et al., 2015;
Culver, 2017). It is worth noting here that outcome-based methods are
indirect, and it may be wise to measure the energy service directly than
its impacts as the same outcome parameters may get influenced by
variables other than the energy service in question.
There are energy poverty measures in literature which combine
different approaches. For instance, the World Bank (2015) has devel­
oped a multi-tier framework (MTF) to go beyond the binary notion of
access and to include various service quantity and quality, affordability,
and pollution parameters, and different tires specify different threshold
values for these parameters.16 The specifications under MTF are com­
bination of different approaches. For example, thresholds on certain tire
on household electric supply specifies a minimum of 800 W peak ca­
pacity (engineering-sense), minimum of 16 h of supply with minimum of
4 h of evening supply (access-sense), maximum spending of 5% of
household income (income-sense), and absence of past accidents and
risks in future (outcome-sense).17 This framework is comprehensive, yet
compromised on simplicity and made the methodology complex and
cumbersome (World Bank, 2015; Culver, 2017). Additionally, the ‘En­
ergy Access Index’ or EAI conceived in this method being an aggregated
measure across all services, turns out to be less meaningful as high­
lighted above (Culver, 2017). Also, any measure that combines different
approaches would suffers from all the limitations highlighted for
different approaches.
3. A new method under access-based approach
3.1. Conceptualizing access as utilization
The proposed method is based on access-based approach to assess
energy poverty. One way of overcoming the limitations of access-based
approach is by considering access as an overall basis of energy utiliza­
tion by the households.18 This is in line with the notion that ‘energy
accessibility’ satisfies the three conditions of ‘energy availability’, ‘en­
ergy affordability’, and ‘energy acceptability’ (Nathan, 2016, 2017). In
other words, if a particular energy is not used by a household, it may be
not available in the first place, or if available, may not be affordable, or if
both available and affordable, but not acceptable to the household. By
following the above approach, one overcomes data limitations as some
of the data on energy utilization become readily available. For example,
in the Indian context, the national sample surveys (NSS) data cover the
household primary source of energy for cooking and lighting as well as
different fuels used in the households (NSSO , 2013). When a household
uses modern fuel, such as LPG or electricity, as its prime source of energy
for cooking, it is implied not only physical connection, but also the
quality and reliability of supply, access to the market, and affordability
of households in the context of both fuel and device. Also, the phrase
‘prime source’ takes care of the situations where a household has access
to multiple sources.
It is important to note that energy poverty based on households’
primary source for certain energy services fits the notion of poverty line
better. When a household substitutes an inefficient fuel by an efficient
16
This comprehensive framework looks into the energy access at three levels:
(i) for basic household services in terms of electricity, cooking, and space-
heating; (ii) for productive activities and enterprises, (iii) for community
facilties in terms of health, education, street lights, government, and public
buildings (World Bank, 2015).
17
There are total five tiers; for specification of each tire under each energy
service, see World Bank (2015).
18
For various notions of energy access, see Pachauri (2011).
H.S.K. Nathan and L. Hari
one as its prime energy source for purposes (such as switching from
kerosene to electricity for lighting) there is a definite jump in the energy
ladder.19
3.2. Focus on cooking and lighting energy services
This paper proposes a new method of computing energy poverty
based on the access-based approach. We define energy poverty on the
basis of an individual’s access to modern forms of energy.20 We have
considered two major aspects of household energy use: cooking and
lighting. The rationale for focussing on these two energy uses comes
from at least four reasons. First, cooking and lighting are of great
importance in today’s households’ energy basket, particularly in the
context of developing countries (Reddy and Nathan, 2013; ADB,
2017).21 Among the different energy needs of a household, cooking and
lighting are the most universal and regular services compared to other
services like space heating and transportation (Wickramasinghe, 2005;
Berkeley National Laboratory, 2009; Greentech, 2010; WHO , 2014).
Being the most basic and the first and the foremost energy use cooking
and lighting constitute not only the main share in household energy
consumption but also form a substantial component in household ex­
penses, particularly for low-income families (Reddy, 2004; CAREPI,
2009; Reddy et al., 2009; Nicholson, 2012).
The second rationale for considering these two energy services is that
lack of access to modern energy sources in cooking and lighting is the
main cause of indoor air pollution causing detrimental effect on health,
mainly that of women’s, leading to premature deaths (Lighting Africa,
2011; Patange et al., 2015; WHO , 2014; Gonz� alez-Eguino, 2015). The
third important reason is that for cooking and lighting energy uses data
are available worldwide. Globally, the International Energy Agency has
been keeping track of people deprived of electricity and modern cooking
fuel through its world energy outlook reports since 2002 (IEA , 2002).
Last but not the least, given the far reaching implications of cooking and
lighting energy services on the quality of life, energy poverty measure­
ment based on households’ primary energy sources for these services
could help target anti-poverty programmes as well programmes specific
to reducing energy deprivation in households (TERI, 2008; Nagothu,
2016).
3.3. Definition of energy poverty and Data Source
Following the access approach and considering the utilization of
modern energy sources for two basic needs of cooking and lighting, we
propose the following definitions of energy poverty:
Energy-nonpoor: An energy-nonpoor is one who belongs to a house­
hold that uses modern fuels as its prime source of energy for both
cooking and lighting and which has no dependency on any energy
inefficient fuels for any purpose whatsoever.
Energy-poor: An energy-poor is one who belongs to a household with
some dependency on any energy inefficient fuels. The energy-poor are
further divided into following three categories.
Transitional energy-poor: A transitional energy-poor is one who be­
longs to a household that uses modern fuels as its prime source of energy
19
Energy ladder concept corresponds to a series of fuel substitution in a
household for different purposes as the economic situation of the household
changes (Hosier and Dowd, 1987; Kowsari and Zerriffi, 2011).
20
The exact fuel types under modern forms of energy considered in this paper
are specified in the section on Study Area and Data Sources.
21
In developed economies, like countries of European Union, access to space
heating/cooling may turn out to be one of the basic needs and the differenti­
ating factor between energy-poor and nonpoor. The same may be the case with
any sub-region of a developing country like India. However, for any such sit­
uation, once the basic needs are identified, the proposed methodology in this
study can be applied to assess energy poverty.
5
Energy Policy 140 (2020) 111397
for both cooking and lighting, but the household has a dependency on
energy inefficient fuels for some purposes.
Moderate energy-poor: A moderate energy-poor is one who belongs to
a household that is deprived of modern fuels as its prime source of en­
ergy for either cooking or lighting, but not in both.
Extreme energy-poor: An extreme energy-poor is one who belongs to a
household that is deprived of modern fuels as its prime source of energy
for both cooking and lighting.
The above-mentioned three sub-groups under energy-poor, viz.,
extreme, moderate, and transitional have no overlaps among them­
selves. As the name suggests, the extreme ones are the poorest in energy
terms, as they are deprived of modern fuels as prime source of energy for
both basic services of cooking and lighting. The moderate ones are the
next sub-group who have such deprivations in only one of the two ser­
vices (not in both). The transitional ones are the least poor among the
three sub-groups in terms of energy as they are not deprived of modern
energy for the two basic services, but their households have some de­
pendency on energy inefficient fuels. When these households cease to
have such dependencies on inefficient fuels, they transit to become en­
ergy-nonpoor.
We propose to compute individuals’ depth and severity of energy
poverty on the basis of dependency on inefficient fuels of the households
to which they belong. Depth of energy poverty corresponds to share of
total household energy consumption catered to by inefficient fuels. This
share indicates the gap from the situation where there is no dependency
on inefficient fuels. Similarly, we consider the square of these shares to
compute severity of energy poverty that accounts for distribution or
inequality among the energy-poor. In the proposed method, we calculate
the average depth and severity of energy poverty by considering only
those who are energy-poor (not for the entire population).
3.4. Novelty of the method
Other than conceptualization of energy access as a combination of
availability, affordability, and acceptability, and considering energy
poverty gap as the fraction of total household energy consumption
coming from inefficient fuels, the novelty of the proposed method occurs
on two counts. First, it grades the energy-poor into different shades:
extreme, moderate, and transitional as opposed to considering all poor
as single homogenous group. As one moves from the extreme energy-
poor status to the energy-nonpoor status dependency on inefficient
fuels reduces. As one moves from extreme to moderate energy-poor, the
household to which the energy-poor belongs sheds its dependency on
inefficient fuels as its prime source for either cooking or lighting. Next,
as one moves from moderate to transitional energy-poor, the household
to which the energy-poor belongs sheds dependency on inefficient fuels
as its prime source for both cooking and lighting. Further, as one up­
grades to the energy-nonpoor status the household sheds its dependency
on any kind of inefficient fuel for any purpose. Unlike the economic
poverty line (that is based on a certain income) or the engineering
approach (that is based on a certain energy quantity), the cut-offs in this
method do not occur along a continuous variable. Rather, they indicate a
definite jump such as a household’s shifting from kerosene to electricity
as its prime source of lighting. Hence, this method suits better to the
practical requirement of categorizing people into different groups
without being as arbitrary as the economic or engineering approaches.
Moreover, the method reduces abruptness by having more than just two
groups: poor and nonpoor.
The second novelty of the method lies in the way it computes the
depth and severity of energy poverty. The conventional measures of
depth (poverty gap ratio, PGR) and severity (squared poverty gap ratio,
SPGR) are improvements over the head count ratio (HCR). PGR assigns
fractional score based on the distance from the poverty threshold and
SPGR further weighs the fractional score by itself to account the
inequality or distribution among the poor (Nathan, 2018; World Bank,
2019). In this sense, the conventional measures of depth and severity are
H.S.K. Nathan and L. Hari
better substitutes of the HCR (Nathan, 2018). These conventional
measures of depth and severity are calculated for the entire population,
where the contribution of nonpoor to the measures is zero (World Bank,
2019). However, the proposed measures of depth and severity of energy
poverty, inspired by Nathan (2018), consider only the energy-poor for
their computations, not the entire population. These two measures,
respectively, are equivalent to income gap ratio (IGR) and squared of
income gap ratio (SIGR) in the income poverty literature (World Bank,
2019). The proposed depth and severity truly reflect the state of the
energy-poor as opposed to a situation in which the values of depth and
severity of energy poverty get diluted with the inclusion of the
energy-nonpoor. So, while the energy poverty HCR is indicative of the
share of the energy-poor in the population, the proposed measures of
depth and severity are indicative of the state of energy poverty of the
energy-poor. In this sense, the proposed measures of depth and severity
are complementary to HCR rather than its substitutes.22
3.5. Study Area and Data Source
The present analysis focusses on India, which is home to the largest
number of people in the world without access to modern cooking fuel or
electricity (IEA, 2017). The latest Census of India (2011) indicates that
approximately two-thirds households use either firewood, crop residue,
dung cake, or charcoal for cooking while approximately one-third of
households are deprived of electricity. According to the IEA (2018)
definition of modern energy, by 2015, India housed 834 million people
without clean cooking fuel facilities and by 2016, 239 million people
were without electricity (IEA, 2017).23
We have focussed on the urban sector for the current study as all
modern forms of energy (electricity, LPG network, etc.) are available in
urban areas (Khandker et al., 2012). Hence, the reason for a household
unable to avail modern forms of energy would be a reflection of its
poverty and unaffordability.24 Moreover, in the urban areas of devel­
oping countries, the nature of energy transition from traditional fuels to
modern fuels has been characterized by the ‘energy ladder’ (signifying
households’ progress to superior fuels with simultaneous abandonment
of inferior fuels) (Kowsari and Zerriffi, 2011). Rural areas, on the con­
trary, show characteristics of ‘energy stacking’ (signifying a ‘partial
switch’ or accumulation of energy options so that households adopting
modern fuels can fall back upon traditional fuels at the time of crisis)
(Kowsari and Zerriffi, 2011).25 The sharply defined energy dependencies
in urban areas are indicative of a definite jump from energy-poor to
nonpoor whereas energy stacking shows more continuous progress.
Also, generally speaking, urban areas in India are characterized by
greater inequalities compared to rural areas (Pal and Ghosh, 2007;
Salve, 2015).
We have considered India’s 35 entities—28 states and seven union
territories (UTs).26 For estimating energy poverty, we have used the NSS
68th Round (NSSO , 2013) Schedule 1.0 data from household con­
sumption expenditure between July 2011 and June 2012. The calcula­
tions are based on the unit level (household level) data on primary
22
For discussion on complementary poverty mesures, see Nathan (2018).
23
The IEA (2012) defines modern energy with respect to clean cooking facility
and electricity connection with minimum consumption (250 kWh in rural areas
and 500 kWh in urban areas).
24
Having said this, we acknowledge that energy poverty measured here
through deprivation of energy services subsumes the unavailability aspects.
25
The crisis may exist either because of supply failure or price fluctuation of
modern energy. For a detailed discussion on the energy transition in developing
economies, see Kowsari and Zerriffi (2011).
26
India currently has 29 states and seven union territories; however, this
study uses data prior to the formation of the newest state, i.e., Telangana, which
was formed out of state of Andhra Pradesh on 2nd June 2014 (India Today,
2014). Thus, for this study, Andhra Pradesh state represents the erstwhile ter­
ritory that included Telangana.
6
Energy Policy 140 (2020) 111397
sources of energy for cooking and lighting and consumption of different
fuels in households for the last 30 days (NSSO , 2013). The different
primary sources of energy for cooking are: coke, coal, charcoal, firewood
and chips, LPG, biogas (gobar gas), dung cake, kerosene, electricity, and
others; and that for lighting are: kerosene, oils other than kerosene, gas,
candle, electricity, and others (NSSO , 2013). Of these, LPG, electricity,
and biogas are considered to be modern sources of energy for cooking.
For lighting, electricity is considered to be a modern energy source. The
fuels that are considered inefficient include coal, coke, charcoal, fire­
wood and chips, dung cake, and kerosene. The calorific values assumed
for different fuels are given in Appendix Table A1.
4. Results and discussion
4.1. Incidence of energy poverty
Fig. 1(a) and (b), respectively, represent the share of the urban
population using different fuels as its prime source of cooking and
lighting. In urban India, overall, more than 70% population uses LPG as
its prime cooking fuel (NSSO , 2013). Firewood and chips and dung
cake, which dominate the rural scene, account for approximately 19%
urbanites’ prime cooking source (NSSO , 2013). The use of such ineffi­
cient fuels is prevalent among slum dwellers (Ahmad, 2014) constituting
17.4% of the urban population (Census of India, 2011).27 These fuels
coupled with poorly built, congested, and overcrowded tenements with
bad ventilation, cause indoor air pollution adversely affecting the health
of women and children (Saksena et al., 2003). Kerosene, a relatively less
polluting fuel than biomass for cooking, is used by 4.5% population as its
prime source for cooking in urban areas (NSSO , 2013). In terms of
lighting, electricity is the predominant medium with 96.3% of the
population using it as its prime source whereas kerosene is used by 3.3%
for the same purpose (NSSO , 2013).
Table A2 presents the incidence of poverty in urban areas of different
entities of India. Fourteen of the 35 entities: 12 states and two UTs –
Lakshadweep, Chhattisgarh, Orissa, Jharkhand, Kerala, Bihar, West
Bengal, Tripura, Uttar Pradesh, Madhya Pradesh, D & N Haveli, Gujarat,
Manipur, and Rajasthan – have more than 30% population deprived of
modern sources of cooking. These are also the same entities with a
greater share of people deprived of modern cooking fuel compared to
the national average which stands at 29.1%. The two entities with less
than 10% urban population deprived of modern cooking fuel are two
northeastern states: Mizoram (5.6%) and Sikkim (9.7%). In terms of
lighting, eight entities – Bihar, Uttar Pradesh, Chhattisgarh, Assam,
Himachal Pradesh, Orissa, Jharkhand, and Tripura – have more than 5%
of urban population deprived of electricity. These eight entities, along
with West Bengal, are the ones with a greater share of its urban popu­
lation deprived of electricity compared to the national average, which
stands at 3.7%. The only two entities with more than 10% population
deprived of modern lighting in urban areas are Bihar (19.4%) and Uttar
Pradesh (11.6%). In contrast, three UTs– A & N Islands, D & N Haveli,
and Lakshadweep – have a 100% population using electricity as its
prime source for cooking. The other three entities with 0.5% or lower
share of its urban population deprived of modern fuels as the prime
source of electricity are: Pondicherry, Nagaland, and Sikkim.
In urban India overall, 3.2% people are extreme energy-poor, 26.3%
moderate energy-poor, 26.1% transitional energy-poor, and 44.3%
energy-nonpoor. Fig. 2 gives the share of the population in the above-
mentioned four groups of energy poverty for different entities of
urban India. There are 14 entities of India housing urban population of
27
As per the Census definition, slum is an area notified or recognized as so by
the state government or identified to be an area with “at least 300 populations
or about 60-70 households of poorly built congested tenements, in unhygienic
environment usually with inadequate infrastructure and lacking in proper
sanitary and drinking waterfacilities.” (GoI, 2010).
H.S.K. Nathan and L. Hari
which more than 30% are either extreme or moderate energy-poor.
These are the same entities with a more than 30% population
deprived of modern sources of cooking. More than half the urban
dwellers in the entities Lakshadweep, Chhattisgarh, and Orissa have
been categorized as extreme or moderate energy-poor. In terms of
extreme energy-poor, Bihar leads the rest with a share of 16.1% followed
by Uttar Pradesh with a corresponding figure of 10.4%. In contrast, four
entities – all UTs including A & N Islands, D & N Haveli, Lakshadweep,
and Pondicherry – have no urban dwellers that may be categorized as
7
Energy Policy 140 (2020) 111397
Fig. 1. Share of population using different
fuels as its prime source of energy for cook­
ing and lighting in urban areas of different
states of India.
Notes: The full name of abbreviation of states
are: AN: Andaman and Nicobar Islands, AP:
Andhra Pradesh, AR: Arunachal Pradesh, AS:
Assam, BR: Bihar, CG: Chhattisgarh, CH:
Chandigarh, DD: Daman and Diu, DL: Delhi,
DN: Dadra & Nagar Haveli, GA: Goa, GJ:
Gujarat, HP: Himachal Pradesh, HR: Har­
yana, JH: Jharkhand, JK: Jammu & Kashmir,
KA: Karnataka, KL: Kerala, ML: Meghalaya,
LD: Lakshadweep, MH: Maharashtra, MN:
Manipur, MP: Madhya Pradesh, MZ: Miz­
oram, NL: Nagaland, OR: Orissa, PB: Punjab,
PD: Pondicherry, RJ: Rajasthan, SK: Sikkim,
TN: Tamil Nadu, TR: Tripura, UK: Uttarak­
hand, UP: Uttar Pradesh, WB: West Bengal.
The entities are sequenced in the decreasing
order of share of modern fuels.
extreme energy-poor. The top three entities in terms of energy-nonpoor
share in urban areas include Delhi, Sikkim, and Haryana. These three
entities contain an energy-nonpoor urban population that is more than
four-fifths. The other two entities with more than two-thirds urban
population energy-nonpoor are Chandigarh (69.4%) and Punjab
(69.3%). Nagaland is the only entity that has more than three-fourths of
its population categorized as transitional energy-poor. It is followed by
another four North-Eastern states, namely, Mizoram, Assam, Manipur,
and Tripura with transitional poor shares as 52.0%, 49.8%, 47.8%, and
H.S.K. Nathan and L. Hari
Fig. 2. Share of energy-poor and within the poor the extreme, moderate, and transitional energy-poor in urban areas of India and its different entities.
Notes: The entities are sequenced in a decreasing order of combined share of extreme and moderate energy-poor. For the full form of the abbreviated names see Notes
of Fig. 1.
46.1%, respectively.
4.2. Observations on incidence of energy poverty
From the incidence of energy poverty in urban areas India and its
different entities, we draw the following two observations. First, the
combined share of extreme or moderate energy-poor is close to the na­
tional average of modern cooking energy deprivation. This indicates
that for those who are deprived of modern energy as their prime source
for lighting are mostly deprived of modern energy as their prime source
of cooking, but not the other way round. From the data it is evident that
of the population that is deprived of electricity, 87% remains deprived of
modern energy sources for cooking in urban areas. In contrast, of those
bereft of modern cooking energy, 11% are deprived of electricity in
urban areas. In this sense, as a greater fundamental need, the depriva­
tion in cooking dictates urban energy poverty in India.
The second observation from the data is that the incidence of energy
poverty is greater for larger sized entities compared to smaller ones.
Ordering the 35 entities of India according to the population in which
Delhi turns out to be the median entity, one may consider the entities
with a greater population than that of Delhi as ‘large’ and entities with
less than the population of Delhi as ‘small’.28 Among the top 10 entities,
in terms of the combined incidence of extreme and moderate energy-
poor in urban areas, eight positions are occupied by large entities. In
contrast, of the bottom 10 positions in the same list, only three positions
are occupied by large entities. One of the reasons for this trend is that for
the same reduction of poverty in terms of percentage points, larger en­
tities require a greater population to be provided with modern energy
services for cooking and lighting.
28
In this way, in the order of population, the 17 large entities are - Uttar
Pradesh, Maharashtra, Bihar, West Bengal, Andhra Pradesh, Madhya Pradesh,
Tamil Nadu, Rajasthan, Karnataka, Gujarat, Odisha, Kerala, Jharkhand, Assam,
Punjab, Chhattisgarh, and Haryana; the 17 small entities are - Jammu and
Kashmir, Uttarakhand, Himachal Pradesh, Tripura, Meghalaya, Manipur,
Nagaland, Goa, Arunachal Pradesh, Puducherry, Mizoram, Chandigarh, Sikkim,
Andaman and Nicobar Islands, Dadra and Nagar Haveli, Daman and Diu, and
Lakshadweep (Census of India, 2011). The large entities together have a share
of 94.6% population and the small entities together account for 4.0% of pop­
ulation, Delhi accounting for the balance (Census of India, 2011).
8
Energy Policy 140 (2020) 111397
4.3. Depth and severity of energy poverty
Table A3 presents the depth and severity of energy poverty of
different entities of India.29 The table gives incidence of energy poverty,
I, which represents the share of the energy-poor population (i.e., the
combined share of extreme, moderate, and transitional energy-poor)
and is equivalent to HCR.30 Both conventional depth and severity (D1
and S1) and newly prosed depth and severity (D2 and S2) are given. D1
and S1 are equivalent to PGR and SPGR, respectively, and D2 and S2 are
equivalent to IGR and SIGR, respectively. Expectedly, D1 and S1, being a
substitute measure of HCR, are bounded by HCR values, whereas D2 and
S2, being complementary measures, are not so constrained (Nathan,
2018). D2 and S2 reflect the poverty situation of the energy-poor more
accurately.
We report ranks of entities as per the conventional and new measures
of depth and severity of energy poverty. We also report the rank dif­
ference when the method changed from conventional to new. The rank
difference is positive (negative) for those whose ranks fell (increased)
under poverty with the new measure suggesting a better situation
compared to the one indicated by the conventional measure. It is worth
noting that the dilution of values in the conventional measures
compared to the new measure will be in proportion to the share of
nonpoor. The country overall, with 44% energy-nonpoor residing in
urban areas, experiences the same level of dilution in values of energy
poverty depth and severity with inclusion of the nonpoor in the con­
ventional measure. So, the values of energy depth and energy severity of
energy-poor (new measure) are substantially greater compared to the
corresponding values of the entire population (conventional measure).
Only for two of the 35 entities the dilution of energy depth and energy
severity with inclusion of energy-nonpoor is less than 10%; whereas, for
as many as 17 entities the dilution is more than 50%. This is reflected in
larger rank differences. Of the 35 entities, for 27 and 24 entities the rank
differences for depth and severity of energy poverty, respectively, are
three or more. Let us elaborate further.
Table 1 and Table 2, respectively, present the entities that evinced
shifts in rankings of 10 or more indicating the degeneration and
improvement in poverty indicated by the severity of energy poverty
29
The entities are ordered as per the new measure of severity of poverty
(SPGR-N). This is so because among the three types of measures considered
here (HCR, PGR, and SPGR), SPGR is the most advanced in the sense that it is
distribution sensitive and satisfies almost all the identified axioms of poverty in
the literature (Zheng, 1997).
30
Note that the ranking of entities as per HCR is the inverse of the ranking as
per the proportion of nonpoor (see Table A2).
H.S.K. Nathan and L. Hari Energy Policy 140 (2020) 111397

Table 1
Entities whose poverty severity worsened as indicated by the new measure compared to the conventional measure.
Entities (States Energy Energy Energy squared Energy Energy squared Rank Rank Rank Rank Rank Rank
or Union poverty head poverty gap poverty gap poverty gap poverty gap PGR-C SPGR- PGR-N SPGR- diff. diff.
Territories) count ratio ratio (conv.) ratio (conv.) ratio (new) ratio (new) C N PGR SPGR
(HCR) (PGR-C) (SPGR-C) (PGR-N) (SPGR-N)

Chandigarh 0.306 0.191 0.148 0.623 0.484 26 24 6 7 20 17

D & N Haveli 0.359 0.251 0.202 0.700 0.562 20 18 2 2 18 16
Rajasthan 0.467 0.287 0.228 0.614 0.488 15 15 7 5 8 10
Haryana 0.181 0.100 0.070 0.556 0.386 33 32 14 22 19 10

Table 2
Entities whose poverty severity improved as indicated by the new measure compared to the conventional measure.
Entities (States or Energy Energy Energy squared Energy Energy squared Rank Rank Rank Rank Rank Rank
Union poverty head poverty gap poverty gap poverty gap poverty gap PGR-C SPGR- PGR-N SPGR- diff. diff.
Territories) count ratio ratio (conv.) ratio (conv.) ratio (new) ratio (new) C N PGR SPGR
(HCR) (PGR-C) (SPGR-C) (PGR-N) (SPGR-N)

Manipur 0.789 0.338 0.258 0.429 0.327 12 12 27 25 15 13

Lakshadweep 0.995 0.590 0.446 0.593 0.448 1 1 9 13 8 12
West Bengal 0.810 0.404 0.322 0.499 0.397 9 9 23 20 14 11
Assam 0.744 0.262 0.198 0.352 0.266 19 19 30 30 11 11
Bihar 0.895 0.450 0.379 0.503 0.423 6 7 21 17 15 10

measure changing from conventional to new.31 The entities in Table 1
typically exhibit a greater share of the energy-nonpoor. With the
exclusion of these nonpoor, the new measures reflect the true poverty
situation of the energy-poor. There is a substantial increase in the values
of energy poverty depth and severity under new measure compared to
those of conventional values. The entities in Table 2 show a typically low
share of the energy-nonpoor. Hence, the new measures reflect a rela­
tively low increase in the values of depth and severity of energy poverty
compared to conventional measures.
Table 3 exhibits some of the contrasting entities that had similar
values of conventional energy severity. Yet, when the new measure was
put in place the ranks changed in opposite directions. The states of Uttar
Pradesh and Madhya Pradesh were ranked 10th and 11th according to
the conventional measure of energy severity. However, when the
nonpoor were excluded in the new measure of severity indicated that the
energy-poor in Madhya Pradesh were worse off compared to their
counterparts in Uttar Pradesh. So, while urban Uttar Pradesh has a
higher share of energy-poor people, the depth and severity of energy
poverty of energy-poor is greater in urban Madhya Pradesh. In other
words, while urban Madhya Pradesh has fewer energy-poor compared to
urban Uttar Pradesh its energy-poor people are poorer.
While calculating the severity of energy poverty for the entire urban
population Uttarakhand and Assam, too, show similar levels. However,
under the new measure, urban Uttarakhand’s depth and severity of
energy poverty compared to that of urban Assam’s is substantially
greater. This shows that while urban Assam has a greater share of the
energy-poor population the poverty level of these energy-poor people, in
terms of depth and severity of energy poverty, is substantially greater for
urban Uttarakhand. A similar trend has been observed for urban Jammu
& Kashmir, which has a greater share of the energy-poor compared to
urban Gujarat; yet urban Gujarat’s energy-poor are poorer than their
counterparts in urban Jammu & Kashmir. So, too, is the case with
Chandigarh that has a lower population share of the energy-poor in
urban areas compared to that of Mizoram’s; yet its energy-poor are
poorer than the latter.
To sum up, we may say, in the conventional measure, the poverty of
the energy-poor gets hidden behind the presence of energy-nonpoor.
Energy-nonpoor having absolutely no dependency on inefficient fuels,
do not contribute to the sum of energy poverty gap (depth) or sum of
31
We limit our discussion to the rank difference in SPGR.
square of energy poverty gap (severity). However, they pull down the
average depth and severity as the total population that divides the sum
includes nonpoor. On the contrary, the proposed measures of depth and
severity, which compute the averages for only the energy-poor (not for
the total population), more accurately reveal the poverty of the energy-
poor.
5. Conclusion and policy implications
As history is often written by the victors so, too, is the history of
energy dominated by energy victors (Birol, 2014). Till recently, energy
poverty remained a neglected area. Acknowledging that one of the
factors responsible for the lag in the context of prioritizing energy
poverty is lack of methodological consensus, this paper visits the major
approaches in literature for assessing energy poverty, viz., economic,
engineering, and access-based, including other approaches. The eco­
nomic and engineering measures of energy poverty have certain draw­
backs including that the former are mere proxy of income poverty and
the latter are too dynamic with the changing technologies and prefer­
ences. Following an access-based approach the study proposes a method
which has novelty in two counts: first, categorization of poor into three
groups instead of considering all the poor as single homogenous lot; and
second, treatment of depth and severity of energy poverty as comple­
mentary measures (not substitutes) to head count ratio and compute the
same for energy-poor (not for the entire population). This method has
been applied for urban areas of different states and union territories of
India using data from the 68th round of NSS. Though this data is
approximately eight years old, it is the latest publicly available national
sample survey data.32 The age of this dataset does not affect the merit of
the paper as the main contribution of the paper is on the methodology of
assessing energy poverty which is applicable to similar datasets.33
The results show the picture of the urban areas in India overall where
the share of extreme energy-poor, moderate energy-poor, transitional
energy-poor, and energy-nonpoor are 3.2%, 26.3%, 26.1%, and 44.3%,
respectively. It has generally been observed that, among the different
entities in India, the larger ones show a higher incidence of energy
32
The next Household Consumer Expenditure survey belongs to 75th round
(2017‒18); the dataset for which is not yet released (The Hindu, 2019).
33
As a future scope, a study on the change in energy poverty between the 68th
and 75th rounds of data can be undertaken, once the latter dataset is made
available.

9
H.S.K. Nathan and L. Hari Energy Policy 140 (2020) 111397

Table 3
Comparison of the entities who had similar energy poverty severity under the conventional measure, but which changed under the new measure.
Entities (States Energy Energy Energy squared Energy Energy squared Rank Rank Rank Rank Rank Rank
or Union poverty head poverty gap poverty gap poverty gap poverty gap PGR-C SPGR- PGR-N SPGR- diff. diff.
Territories) count ratio ratio (conv.) ratio (conv.) ratio (new) ratio (new) C N PGR SPGR
(HCR) (PGR-C) (SPGR-C) (PGR-N) (SPGR-N)

Madhya Pradesh 0.576 0.353 0.292 0.612 0.507 11 11 8 4 3 7

Uttar Pradesh 0.752 0.386 0.317 0.514 0.421 10 10 20 19 10 9
Uttarakhand 0.434 0.242 0.198 0.558 0.456 21 20 13 11 8 9
Assam 0.744 0.262 0.198 0.352 0.266 19 19 30 30 11 11
Gujarat 0.412 0.229 0.182 0.554 0.443 22 21 15 14 7 7
Jammu & 0.560 0.267 0.181 0.478 0.323 18 22 25 26 7 4
Kashmir
Chandigarh 0.306 0.191 0.148 0.623 0.484 26 24 6 7 20 17
Mizoram 0.584 0.201 0.141 0.344 0.241 24 25 31 31 7 6

poverty. Additionally, the data shows that between the two services –
cooking and lighting – it is the former that dictates energy poverty. This
is because one who is deprived of modern lighting fuels is most likely to
be deprived of modern cooking fuels, but not the other way round. This
shows up cooking as a more fundamental need.
The policy implications of the new measures are multitude. The
study’s proposed method of finding depth and severity of energy poverty
has the capability of highlighting the poverty level of the energy-poor.
As a complementing measure to head count ratio these measures
would direct the attention of pro-poor energy programmes to the regions
where there is greater level of poverty of the poor. It is made evident in
this study that regions with higher proportion of energy poor may not be
same with greater depth and severity of energy poverty in the light of the
new measures. For instance, urban Gujarat has a smaller proportion of
energy-poor people compared to Uttar Pradesh or Bihar. However, it
may be readily seen that Gujarat’s energy-poor are worse off compared
to their counterparts in Uttar Pradesh and Bihar. In other words while
urban Bihar or urban Uttar Pradesh have a greater population share in
the energy-poor category, people belonging to this category are poorer
in urban Gujarat.
Value judgement is based on the choice of a measure and in this
sense, measures have considerable bearing on policy choices (Lipton and
Ravallion, 1993; Moore, 2012). The conventional measures of depth and
severity are bounded by HCR. A large share of the nonpoor tends to
substantially lower the average values of depth and severity of energy
poverty. This reduces the policy sensitivity capacity of the measures.34
However, the new measure, by excluding energy-nonpoor, has been able
to capture exclusively the poverty level of energy-poor and thereby can
signal the policies to reduce the same. The policy implications are not
limited to only energy sector. The depth and severity of energy poverty
of a household which is based on its extent of dependency on inefficient
fuels are indicative of not only poor energy services, but also
energy-induced emissions, and emission-induced ill-health. To sum up,
the new measures of depth and severity are useful in terms of depicting
the level of poverty of the poor without being bound by HCR values.
Hence, these complementary measures to HCR must be given promi­
nence in energy policy discourse.
Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
CRediT authorship contribution statement
Hippu Salk Kristle Nathan: Conceptualization, Methodology,
Formal analysis, Validation, Investigation, Resources, Writing - original
draft, Writing - review & editing. Lakshmikanth Hari: Conceptualiza­
tion, Methodology, Software, Formal analysis, Investigation, Resources,
Data curation, Writing - review & editing.

Appendix

Table A1
Fuels and their calorific values

Fuel Calorific values and prices

Cokea 24.8 million Btu/ton ⇒ 28842.4 kJ/kg

Firewood and chipsb 4500 kcal/kg ⇒ 18828 kJ/kg
Electricity (std. unit) 3600 kJ/kg
Dung cakec,d 2092 kcal/kg ⇒ 8752.9 kj/kg; (Price ¼ Rs. 3/kg) ⇒2917.643 kJ/Rs.
Keroseneb,e 10638 kcal/kg ⇒44509.4 kJ/kg; (density ¼ 8724.2 kcal/lit) ⇒ 36502.153 kJ/lit
Coalb 4000 kcal/kg ⇒ 16736 kJ/kg
LPGb 11300 kcal/kg ⇒ 47279.2 kJ/kg
Charcoalf 6900 kcal/kg ⇒28869.6 kJ/kg
Gobar gasg,h 4713 kcal/m3 ⇒ 19719.2 kj/m3; (Price ¼ Rs. 6.87/m3) ⇒ 2870.334 kJ/Rs.
Notes: The calorific values are given in units that comply with the data collected in the NSS survey. For instance, for
dung cake, data is collected in terms of household expenditure in Rs.; hence the calorific value is expressed in terms

34
This argument is similar to policy sensitivity arguments put forward by Nathan and Mishra (2017) in the context of group differential and by Subramanian (2004)
in the context of literacy. The former argues that the group differential measures must be greater than simple difference and the latter suggested to extend the
measure of effective literacy based on proximate and secluded illiterates as proposed by Basu and Foster (1998), so as to make the measure a lower compared to the
conventional literacy rate.

10
H.S.K. Nathan and L. Hari Energy Policy 140 (2020) 111397

of kJ/Rs. The references for the data is given below.

a
Calorific value from Speight (2013).
b
Calorific value from TERI (2007).
c
Calorific value from KVIC (1983) as reported in Sampath Kumar et al. (1985).
d
Price from GoI (2015); Lal et al. (2016).
e
Density of kerosene, 0.8201 g/cm3 as reported in Endmemo (2015).
f
Calorific value from Ramachandra (2000).
g
Calorific value from Ramachandra (2000).
h
Price of biogas was obtained from inflation-adjusted current price (Chaba, 2018; RBI , 2017).

Table A2Incidence of energy poverty in urban areas of different states of India

States Deprived of Deprived of Energy-poor Energy- Ext. or mod. Rank Rank Rank Rank Rank
modern modern nonpoor energy-poor EþM E M T N
Extreme Moderate Transitional
cooking fuel lighting fuel (N) (%) (E þ M) (%)
(E) (%) (M) (%) (T) (%)
(%) (%)

Lakshadweep 57.6 0.0 0.0 57.6 41.8 0.5 57.6 1 12 3 35 35

Chhattisgarh 56.4 7.0 6.2 51.0 12.9 29.9 57.2 2 4 20 34 24
Orissa 53.2 6.4 6.3 46.9 23.7 23.1 53.2 3 6 19 32 27
Jharkhand 47.3 5.2 4.2 44.1 32.1 19.6 48.3 4 1 32 31 29
Bihar 43.0 19.4 16.1 30.1 43.2 10.5 46.2 5 2 30 30 33
Kerala 43.7 2.3 1.9 42.3 41.1 14.7 44.1 6 10 6 29 31
West Bengal 42.7 4.3 3.9 39.3 37.9 19.0 43.1 7 32 1 26 30
Tripura 41.5 5.1 4.0 38.6 46.1 11.3 42.6 8 11 7 28 32
Uttar Pradesh 36.6 11.6 10.4 27.5 37.3 24.8 37.8 9 7 18 25 26
Madhya 36.4 1.8 1.6 34.9 21.1 42.4 36.6 10 9 9 27 21
Pradesh
D & N Haveli 31.1 0.0 0.0 31.1 4.7 64.1 31.1 11 20 2 23 8
Gujarat 31.0 2.1 2.0 29.1 10.1 58.8 31.1 12 8 12 24 11
Manipur 30.9 1.9 1.7 29.3 47.8 21.1 31.1 13 24 4 19 28
Rajasthan 30.4 1.9 1.9 28.5 16.3 53.3 30.4 14 19 5 21 14
Daman & Diu 28.3 1.9 1.9 26.4 19.8 51.9 28.3 15 3 27 15 16
Karnataka 27.0 1.0 1.0 26.0 31.2 41.8 27.0 16 18 10 20 22
Meghalaya 24.4 2.5 0.9 25.0 24.0 50.0 26.0 17 29 8 14 17
Chandigarh 25.5 1.7 1.7 23.8 5.2 69.4 25.5 18 17 13 18 4
Assam 24.0 6.7 6.0 18.6 49.8 25.6 24.6 19 16 16 12 25
Himachal 20.1 6.4 3.2 20.2 11.6 65.0 23.4 20 22 11 13 7
Pradesh
Tamil Nadu 22.7 1.2 1.0 21.9 28.3 48.8 22.9 21 5 29 11 18
Maharashtra 22.5 1.0 0.9 21.7 24.3 53.1 22.6 22 27 14 16 15
Uttarakhand 22.4 3.2 3.1 19.3 21.0 56.6 22.5 23 21 15 8 12
A & N Islands 22.1 0.0 0.0 22.1 11.9 66.0 22.1 24 26 17 9 6
Jammu & 19.0 1.8 1.1 18.6 36.3 44.0 19.7 25 15 22 22 20
Kashmir
Punjab 17.7 1.1 1.1 16.6 13.0 69.3 17.7 26 14 23 10 5
Andhra 17.3 1.3 0.9 16.7 26.4 55.9 17.6 27 13 24 6 13
Pradesh
Arunachal 13.6 3.3 1.3 14.2 36.5 47.9 15.6 28 23 21 7 19
Pradesh
Pondicherry 13.9 0.4 0.0 14.3 25.1 60.5 14.3 29 30 25 1 10
Delhi 11.3 1.5 0.5 11.8 2.9 84.9 12.3 30 31 26 4 1
Nagaland 11.3 0.4 0.4 10.9 79.5 9.2 11.3 31 33 28 2 34
Haryana 10.6 1.2 0.6 10.7 6.8 81.9 11.3 32 34 31 5 3
Goa 11.2 2.0 2.0 9.2 27.5 61.3 11.2 33 25 33 3 9
Sikkim 9.7 0.5 0.5 9.2 6.2 84.1 9.7 34 35 34 17 2
Mizoram 5.5 1.1 0.2 6.2 52.0 41.6 6.4 35 28 35 33 23
India 29.1 3.7 3.2 26.3 26.1 44.3 29.5
Note: The entities are sequenced in decreasing order of combined share of extreme and moderate energy-poor (E þ M).

Table A3Depth and severity of energy poverty in urban areas of different states of India

States Incidence of Depth of Severity of Depth of Severity of Rank Rank Rank Rank Rank Rank
energy poverty energy poverty Energy poverty energy poverty Energy poverty D1 S1 D2 S2 diff. D2- diff.
(I) (conv.) (D1) (conv.) (S1) (new) (D2) (new) (S2) D1 S2–S1

Chhattisgarh 0.701 0.500 0.437 0.713 0.622 4 2 1 1 3 1

D & N Haveli 0.359 0.251 0.202 0.700 0.562 20 18 2 2 18 16
Orissa 0.769 0.484 0.418 0.630 0.544 5 4 5 3 0 1
Madhya 0.576 0.353 0.292 0.612 0.507 11 11 8 4 3 7
Pradesh
Rajasthan 0.467 0.287 0.228 0.614 0.488 15 15 7 5 8 10
Kerala 0.853 0.538 0.414 0.631 0.486 3 5 4 6 1 1
Chandigarh 0.306 0.191 0.148 0.623 0.484 26 24 6 7 20 17
Arunachal 0.521 0.308 0.251 0.592 0.482 14 13 10 8 4 5
Pradesh
(continued on next page)

11
H.S.K. Nathan and L. Hari Energy Policy 140 (2020) 111397

(continued )
States Incidence of Depth of Severity of Depth of Severity of Rank Rank Rank Rank Rank Rank
energy poverty energy poverty Energy poverty energy poverty Energy poverty D1 S1 D2 S2 diff. D2- diff.
(I) (conv.) (D1) (conv.) (S1) (new) (D2) (new) (S2) D1 S2–S1

Nagaland 0.908 0.590 0.436 0.650 0.481 2 3 3 9 1 6

Jharkhand 0.804 0.441 0.375 0.549 0.466 7 8 17 10 10 2
Uttaranchal 0.434 0.242 0.198 0.558 0.456 21 20 13 11 8 9
Daman & Diu 0.481 0.278 0.217 0.579 0.452 17 16 11 12 6 4
Lakshadweep 0.995 0.590 0.446 0.593 0.448 1 1 9 13 8 12
Gujarat 0.412 0.229 0.182 0.554 0.443 22 21 15 14 7 7
Tripura 0.887 0.437 0.384 0.492 0.433 8 6 24 15 16 9
Meghalaya 0.500 0.280 0.216 0.560 0.433 16 17 12 16 4 1
Bihar 0.895 0.450 0.379 0.503 0.423 6 7 21 17 15 10
Karnataka 0.582 0.319 0.245 0.549 0.422 13 14 16 18 3 4
Uttar Pradesh 0.752 0.386 0.317 0.514 0.421 10 10 20 19 10 9
West Bengal 0.810 0.404 0.322 0.499 0.397 9 9 23 20 14 11
Punjab 0.307 0.167 0.122 0.546 0.396 30 30 18 21 12 9
Haryana 0.181 0.100 0.070 0.556 0.386 33 32 14 22 19 10
A & N Islands 0.340 0.179 0.128 0.526 0.375 27 27 19 23 8 4
Himachal 0.350 0.175 0.127 0.500 0.362 29 28 22 24 7 4
Pradesh
Manipur 0.789 0.338 0.258 0.429 0.327 12 12 27 25 15 13
Jammu & 0.560 0.267 0.181 0.478 0.323 18 22 25 26 7 4
Kashmir
Tamil Nadu 0.512 0.221 0.151 0.432 0.294 23 23 26 27 3 4
Maharashtra 0.469 0.194 0.136 0.413 0.290 25 26 28 28 3 2
Andhra 0.441 0.178 0.126 0.404 0.286 28 29 29 29 1 0
Pradesh
Assam 0.744 0.262 0.198 0.352 0.266 19 19 30 30 11 11
Mizoram 0.584 0.201 0.141 0.344 0.241 24 25 31 31 7 6
Pondicherry 0.395 0.120 0.071 0.305 0.181 31 31 32 32 1 1
Goa 0.387 0.111 0.067 0.288 0.173 32 33 33 33 1 0
Delhi 0.151 0.038 0.025 0.249 0.163 34 34 34 34 0 0
Sikkim 0.159 0.027 0.019 0.173 0.118 35 35 35 35 0 0
India 0.557 0.282 0.220 0.506 0.395
Note: The entities are sequenced in decreasing order of proposed severity of energy poverty (S2).

References Culver, L.C., 2017. Energy Poverty: what You Measure Matters, White Paper for
Symposium on Reducing Energy Poverty with Natural Gas: Changing Political,
Business, and Technology Paradigms. Stanford University, CA.
Acharya, R.H., Sadath, A.C., 2019. Energy poverty and economic development:
DTI (Dept. Of Trade and Industry), 2006. Energy—its Impact on the Environment and
household-level evidence from India. Energy Build. 183, 785–791.
Society. DTI Report. DTI, Government of UK.
ADB (Asian Development Bank), 2007. Energy for All: Addressing the Energy,
Easterly, W., 2009. How the Millennium development goals are unfair to Africa. World
Environment, and Poverty Nexux in Asia. April.
Dev. 37 (1), 26–35.
ADB, 2017. ‘Myanmar Energy Consumption Surveys’, September.
Endmemo, 2015. Density of Common Liquids accessed July 18, 2018. http://www.
Aguilar, J.M., Ramos-Real, F.J., Ramírez-Díaz, A.J., 2019. Improving indicators for
endmemo.com/cconvert/kgl.php.
comparing energy poverty in the canary Islands and Spain. Energies 12 (11), 1–15.
Foster, V., 2000. “Measuring the Impact of Energy Reform-Practical Options”. Viewpoint,
Ahmad, S., 2014. “Tackling the Switch from Traditional Cooking Fuels in Indian Slums”,
Note#210, Private Sector and Infrastructure Network. The World Bank Group.
Our Word. United Nations University.
Foster, V., Jean-Philippe, T., Wodon, Q., 2000. Energy prices, energy efficiency, and fuel
Alkire, S., Foster, J., 2011. Counting and multidimensional poverty measurement.
poverty, Latin America and Caribbean Regional Studies Program. The World Bank,
J. Publ. Econ. 95 (7–8), 476–487.
Washington, D.C.
Bachao, Bijli, 2018. Domestic Electricity LT Tariff Slabs and Rates for All States in India
Gillingham, K., Rapson, D., Wagner, G., 2014. “The Rebound Effect and Energy
in 2018.
Efficiency Policy”, Discussion Paper, RFF DP 14-39, Resources for the Future.
Barnes, D.F., Hussain, S., Banerjee, S.G., 2014. The development impact of energy access.
GoI (Govt of India), 2010. Report of the Committee on Slum Statistics/Census. National
In: Halff, A., Sovacool, K., Rozhon, J. (Eds.), Energy Poverty: Global Challenges and
Building Organization, Ministry of Housing and Urban Poverty Alleviation, GoI, New
Local Solutions. Oxford University Press, Oxford, pp. 54–73.
Delhi, p. 8.
Basu, K., Foster, J.E., 1998. On measuring literacy. Econ. J. 108 (6), 1733–1749.
GoI, 2015. Samarthya, Technical Training Manual (MGNREGA). Ministry of Rural
Berkeley National Laboratory, 2009. Residential and Transport Energy Use in India: Past
Development, GoI, March.
Trend and Future Outlook. Environmental Energy Technologies Division, Ernest
Goldemberg, J., Johansson, T.B., 1995. Energy as an Instrument for Socio-Economic
Orlando Lawrence, LBNL-1753E.
development. Tech. Rep.. United Nations Development Programme, New York.
Birol, F., 2007. Energy economics: a place for energy poverty in the Agenda? Energy J.
Goldemberg, J., Johansson, T.B., Reddy, A.K.N., Williams, R.H., 1985. Basic needs and
28 (3), 1–6.
much more with one kilowatt per capita. Ambio 14 (4–5), 190–200.
Birol, F., 2014. Achieving energy for all will not cost the earth. In: Halff, A., Sovacool, K.,
Gonz� alez-Eguino, M., 2015. Energy poverty: an overview. Renew. Sustain. Energy Rev.
Rozhon, J. (Eds.), Energy Poverty: Global Challenges and Local Solutions. Oxford
47, 377–385.
University Press, Oxford, pp. 11–20.
Goozee, H., 2017. Energy, Poverty and Development: a Primer for the Sustainable
Boardman, B., 1991. Fuel Poverty: From Cold Homes to Affordable Warmth. Belhaven
Development Goals. International Policy Centre for Inclusive Growth, Brasilia. ISSN
Press, London.
1812-108x, Working Paper# 156.
Bradsahw, J.S., Middleton, S., Davis, A., Oldfield, N., Smith, N., Cusworth, L.,
Greentech, 2010. Solar Water Heaters in India: Market Assessment Studies and Surveys
Williams, J., 2008. A Minimum Income Stardard for Britain. What People Think? ”
for Different Sectors and Demand Segments. Greentech Knowledge Solutions Pvt.
Joseph Rowntree Foundation, York, UK.
Ltd., New Delhi.
CAREPI, 2009. Identification of energy-related needs and priorities of poor communities
Halff, A., Sovacool, K., Rozhon, J., 2014. “Introduction: the end of energy poverty –
in karangmojo, yogyakarta”, version 1: 3/23/2009. A Report of Work Package 2. In:
pathways to Development”. In: Halff, A., Sovacool, K., Rozhon, J. (Eds.), Energy
Rosyidi, A.P., Setiartiti, L., Lesmana, S.B., Saibun (Eds.). March.
Poverty: Global Challenges and Local Solutions. Oxford University Press, Oxford,
Census of India, 2011. Office of the Registrar General & Census Commissioner. Govt. of
pp. 4–10.
India. http://censusindia.gov.in/2011-Common/CensusData2011.html.
Hills, J., 2012. Getting the Measure of Fuel Poverty. CASE Report #72. Center for
Chaba, A.A., 2018. Waste to Wealth: for These Families, LPG Is Now Alternate Fuel. The
Analysis of Social Exclusion. ISSN 1465-3001, March.
Indian Express. March 1.
Hosier, R.H., Dowd, J., 1987. Household fuel choice in Zimbabwe: an empirical test of
Commonwealth Secretariat, 2014. Istanbul Programme of Action for the LDCs (2011-
the energy ladder hypothesis. Resour. Energy 9 (4), 347–361.
2020): Monitoring Deliverables, Tracking Progress – Analytical Perspective.
IAEA (International Atomic Energy Authority), 2008. Integrated Energy Planning for
Commonwealth Secretariat.
Sustainable Development. IAEA, January.

12
H.S.K. Nathan and L. Hari
IEA (International Energy Agency), 2002. World Energy Outlook 2002. OECD/IEA, Paris.
IEA, 2012. World Energy Outlook 2012. OECD/IEA, Paris.
IEA, 2014. Sustainable Energy for All: Global Tracking Framework Report, IEA and the
World Bank. World Bank, Washington DC.
IEA, 2017. Energy Access Outlook 2017: from Poverty to Prosperity. OECD/IEA, Paris.
IEA, 2018. World Energy Model Documentation: 2018 Version. OECD/IEA, Paris.
Available at: https://www.sipotra.it/wp-content/uploads/2019/03/WORLD-EN
ERGY-MODEL-DOCUMENTATION-2018.pdf. accessed on 08 Dec 2019.
Jain, A., Shahidi, T., 2019. Guiding action: a user centric approach to define, Measure,
and Manage Electricity Access. Econ. Energy Environ. Policy 8 (1), 19–32.
Kemmler, A., 2007. Characterisitcs of Energy Access to the Poor and to Impoverished
Regions in India. Swiss Federal Institute of Technology (ETH), Zürich. Dissertation
17033, ETDE-CH-10071.
Khandker, S.R., Barnes, D.F., Samad, H.A., 2012. Are the energy poor also income poor?
Evidence from India. Energy Policy 47, 1–12.
Kowsari, R., Zerriffi, H., 2011. Three dimensional energy profile: a conceptual
framework for assessing household energy use. Energy Pol. 39 (12), 7505–7517.
Kumar, T.S., Vasudevan, P., Patwardhan, S.V., 1985. Pattern of non-commercial energy
consumption and availability in the Indian domestic sector- A case study. Agric.
Wastes 12 (1), 55–60.
KVIC (Khadi and Village Industries Commission), 1983. Gobar Gas: Why and How. KVIC,
Bombay.
Lal, R.M., Nagpure, A.S., Luo, L., Tripathi, S.N., Ramaswami, A., Bergin, M.H., Russell, A.
G., 2016. “Municipal solid waste and dung cake burning: discoloring the Taj Mahal
and human health impacts in Agra”, Supplementary material. Environ. Res. Lett. 11
(10) https://doi.org/10.1088/1748-9326/11/10/104009.
Leach, G., 1987. Household energy in south Asia. Biomass 12 (3), 155–184.
Lipton, M., Ravallion, M., 1993. “Poverty and Policy”, Working Paper (WPS 1130).
Policy Research Department, The World Bank.
Lighting Africa, 2011. Expanding women’s role in Africa’s modern off-grid lighting
market”, authored. In: Alstone, P., Niethammer, C., Mendonça, B., Eftimi, A. (Eds.),
October 4, Lighting Africa. The World Bank.
Madureira, N.L., 2014. Key Concepts in Energy. Springer, London, p. 10.
Maxim, A., Mihai, C., Apostoaie, C.-M., Popescu, C., Istrate, C., Bostan, I., 2016.
Implications and measurement of energy poverty across the European union.
Sustainability 8 (5), 1–20, 483.
Mirza, B., Szirmai, A., 2010. Towards a New Measurement of Energy Poverty: A Cross-
Community Analysis of Rural Pakistan, #2010-024, Working Paper Series, UNU-
MERIT. United Nations University.
Modi, V., McDade, S., Lallement, D., Saghir, J., 2005. Energy Services for the Millennium
Development Goals. World Bank Energy Sector Management Assistance Program and
United Nations Development Programme, Washington, DC and New York.
Moore, R., 2012. Definitions of fuel poverty: implications for policy. Energy Policy vol.
49 (C), 19–26.
Nagothu, S., 2016. Measuring Multidimensional Energy Poverty: the Case of India.
Master Thesis. Norwegian School of Economics, Spring.
Nathan, H.S.K., 2016. Energy Access and Energy Poverty: Concepts and Insights on
Measurements, Asian Development Bank Institute Seminar Series, 31 August. ADBI.
Nathan, H.S.K., 2017. “Energy Poverty: Concepts, Measurement and Impacts on Quality
of Life”, LSE India Observatory Seminar, 10 May.
Nathan, H.S.K., 2018. Substitutes and Complements – the Curious Case of Poverty
Measure. NCDS, September. Working Paper# 71.
Nathan, H.S.K., Mishra, S., 2017. Group differential for attainment and failure. J. Int.
Dev. 29 (4), 520–532.
Nathan, H.S.K., Raj, B., 2016. “Can energy equality improve India’s human development
index?”, 11. Indian Economist, December.
Nathan, H.S.K., Rajendra, Mishra, S., 2015. Towards measuring interventions in rainfed
areas. In: Devi Prasad, J., Gangaiah, B., Suman Chandra, K. (Eds.), Agricultural Risk
Management. PharmaMed Press/BS Publications, Hyderabad, pp. 274–293 (Chapter
17).
Nicholson, D., 2012. Household Energy Market Assessment: an Assessment of Household
Energy Use and Supply in Mandalay, Chin, and Rakhine States, Myanmar. Mercy
Corps Myanmar.
NSSO (National Sample Survey Organization), 2013. Key Indicators of Household
Consumption Expenditures in India, 68th Round (July 2011 to June 2012), NSSO,
Ministry of Statistics and Programme Implementation. Gov. of India.
Nussbaumer, P., Bazilian, M., Modi, V., 2012. Measuring energy poverty: focusing on
what matters. Renew. Sustain. Energy Rev. vol. 16 (1), 231–243.
Pachauri, S., 2011. Reaching an international consensus on defining modern energy
access. Curr. Opin. Environ. Sustain. vol. 3 (4), 235–240.
Pachauri, S., Spreng, D., 2004. Energy use and energy access in relation to poverty. Econ.
Polit. Wkly. 39 (3), 271–278.
Pachauri, S., Muller, A., Kemmler, A., Spreng, D., 2004. On measuring energy poverty in
Indian households. World Dev. vol. 32 (12), 2083–2104.
Pal, P., Ghosh, J., 2007. Inequality in India: A Survey of Recent Trends. Department of
Economic and Social Affairs, United Nations, New York. DESA Working Paper# 45.
Patange, O.S., Ramanathan, N., Rehman, I.H., Tripathi, S.N., Mishra, A., Kar, A.,
Graham, E., Singh, L., Bahadur, R., Ramanathan, V., 2015. Reductions in indoor
13
Energy Policy 140 (2020) 111397
black carbon concentrations from improved biomass stoves in rural India. Environ.
Sci. Technol. 49 (7), 4749–4756.
Ramachandra, T.V., 2000. Energy Alternatives: Renewable Energy and Energy
Conservation Technologies, Calorific Value of Different Fuels (Reference Section).
Energy Research Group, Center of Ecological Sciences accessed 18 Jul 2018. http
://wgbis.ces.iisc.ernet.in/energy/paper/alternative/energyalternative.html.
RBI (Reserve Bank of India), 2017. Handbook of Statistics on the Indian Economy,
Table 41, Mumbai: RBI, p. 91. September 15.
Reddy, A.K.N., 1994. Energy and social issues. Econ. Polit. Wkly. 34 (49), 3435–3445.
Reddy, A.K.N., 1999. Goals, strategies and policies for rural energy. Econ. Polit. Wkly. 34
(49), 3435–3445.
Reddy, B.S., 2004. “Technological Substitution and Changes in the Household Energy
Consumption Pattern: A Case Study of India”, 2004 ACEEE Summer Study on Energy
Efficiency in Buildings–Proceedings. European Council for an Energy Efficient
Economy.
Reddy, B.S., Nathan, H.S.K., 2013. “Energy in the development strategy of Indian
households—the missing half”. Renew. Sustain. Energy Rev. 18, 203–210.
Reddy, B.S., Balachndra, P., Nathan, H.S.K., 2009. “Universalization of access to modern
energy services in Indian households—economic and policy analysis”. Energy Pol. 37
(11), 4645–4657.
Sadath, A.C., Acharya, R.H., 2017. Assessing the extent and intensity of energy poverty
using Multidimensional Energy Poverty Index: empirical evidence from households
in India. Energy Pol. 102, 540–550.
Saksena, S., Singh, P.B., Prasad, R.K., Prasad, R., Malhotra, P., Joshi, V., Patil, R.S., 2003.
Exposure of infants to outdoor and indoor air pollution in low-income urban areas - a
case study of Delhi. J. Expo. Anal. Environ. Epidemiol. 13 (3), 219–230.
Salve, P., 2015. India’s wealth is rising. So is inequality accessed October 3, 2015. htt
p://scroll.in/article/699519/indias-wealth-is-rising-so-is-inequality.
Schuessler, R., 2014. “Energy Poverty Indicators: Conceptual Issues—Part I: The Ten-
Percent-Rule and Double Median/Mean Indicators”, ZEW Discussion Papers# 14-
037, Zentrum für Europ€ aische Wirtschaftsforschung (ZEW), Mannheim.
SE4All (Sustainable Energy for All), 2014. Annual Report. SE4All.
Sovacool, B., 2014. Defining, measuring, and tackling energy poverty. In: Halff, A.,
Sovacool, K., Rozhon, J. (Eds.), Energy Poverty: Global Challenges and Local
Solutions. Oxford University Press, Oxford, pp. 21–53.
Speight, J.G., 2013. The Chemistry and Technology of Coal, third ed. CRC Press (Taylor
& Francis Group, Boca Raton (Florida), p. 518.
Subramanian, S., 2004. Measuring literacy: some extensions of the Basu–Foster
framework. J. Dev. Econ. 73 (1), 453–463.
Swan, L.G., Ugursal, V.I., 2009. Modeling of end-use energy consumption in the
residential sector: a review of modeling techniques. Renew. Sustain. Energy Rev. 13
(8), 1819–1835.
TERI (The Energy Research Institute), 2007. TERI Energy Data Directory and Yearbook
(TEDDY) 2005-06. TERI Press, New Delhi, p. 363.
TERI, 2008. Supply of Clean Energy Services to Urban and Peri-Urban Poor’, Final
Report, Global Network on Energy for Sustainable Development. TERI, New Delhi.
Project Report No 2007UD21.
The Hindu, 2019. “Protest Statement for Release of NSSO Consumption Survey Data”,
Business Line. The Hindu. November 18.
India Today, 2014. Telangana: did you know how India got its 29 states and 7 UTs? A.D.
Ashok. June 02.
UN (United Nations), 2005. “The Energy Challenge for Achieving the Millennium
Development Goals”, UN-energy.
UN, 2010. Delivering on Energy: an Overview of Activities by UN-energy and its
Members. UN-Energy.
UN, 2011. 2012 International Year of Sustainable Energy for All accessed May 17, 2018.
https://www.un.org/en/events/sustainableenergyforall.
UN, 2012. United Nations General Assembly Declares 2014-2024 Decade of Sustainable
Energy for All accessed May 17, 2018. https://www.un.org/press/en/2012/
ga11333.doc.htm.
UN, 2013. “Our common action Agenda to promote sustainable energy in the least
developed countries”, Co-chairs communication. High-Level Event on Sustainable
Energy for the Least Developed Countries. New York, 23 September.
UNDP, 2009. The Energy Access Situation in Developing Countries, A Review Focusing
on Least Developed Countries and Sub-saharan Africa. UNDP and World Health
Organization, New York, USA.
UNDP, 2015. Sustainable Development Goals. UNDP.
Watts, H., 1968. An economic definition of poverty. In: Moynihan, D.P. (Ed.), On
Understanding Poverty. Basic Books, New York.
WHO (World Health Organization), 2014. Indoor Air Quality Guidelines: Household Fuel
Combustion. WHO, Geneva.
Wickramasinghe, A., 2005. Gender, Modern Biomass Energy Technology and Poverty:
Case Study in Sri Lanka. ENERGIA and DFID KaR Research Project, p. R8346.
World Bank, 2015. Beyond Connections: Energy Access Redefined, Technical Report
008/15, Energy Sector Management Assistance Programme. World Bank.
World Bank, 2019. Technical Note: Measuring Poverty and Analyzing Changes in Poverty
over Time. World Bank. Available at: http://siteresources.worldbank.org/INTPA/Re
sources/tn_measuring_poverty_over_time.pdf. accessed on 09 Dec 2019.
Zheng, B., 1997. Aggregate poverty measures. J. Econ. Surv. 11 (2), 123–162.