Accepted Manuscript Fuel Switching in Slum and Non-slum Households in Urban India Sohail Ahmad, Jose A. Puppim de Oliveira PII:
S0959-6526(15)00076-1
DOI:
10.1016/j.jclepro.2015.01.072
Reference:
JCLP 5149
To appear in:
Journal of Cleaner Production
Received Date: 4 April 2014 Revised Date:
21 January 2015
Accepted Date: 21 January 2015
Please cite this article as: Ahmad S, Puppim de Oliveira JA, Fuel Switching in Slum and Non-slum Households in Urban India, Journal of Cleaner Production (2015), doi: 10.1016/j.jclepro.2015.01.072. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Fuel Switching in Slum and Non-slum Households in Urban India Ahmad, Sohaila, b, 1 and Puppim de Oliveira, Jose A.a
Graduate School of Decision Science and Technology, Tokyo Institute of Technology, Tokyo, Japan
SC
b
United Nations University Institute for the Advanced Study of Sustainability (UNU-IAS) 5–53–70 Jingumae, Shibuya-ku, Tokyo 150-8925 Japan
RI PT
a
Abstract
TE D
M AN U
Improving access to modern fuels is essential in developing countries in reducing adverse human health and environmental impacts caused by traditional fuels. Fuels use in developing countries is heterogeneous across households. This paper estimates drivers of fuel switching in non-slum and slum households in urban India, using a discrete choice model on a nationally representative micro data. The choices considered are three categories of cooking fuels: traditional – firewood, dung, crop residue and coal/charcoal; modern – kerosene and liquefied petroleum gas (LPG); and mixed fuels. The results suggest that the patterns of fuels use are consistent with the energy ladder theory in urban India. In addition to income, the major determinants of modern fuels uptake are fuels prices, access to electricity and supply water, and education attainment. The increasing price of LPG affects the low-income non-slum and the high-income slum households negatively. The analyses make a strong case for applying differential subsidies on modern fuels employing multidimensional aspects of poverty. Moreover, there is a need for partial diversion of existing fuels subsidies on improving physical and social capitals, which will result in uptake of modern fuels, particularly among disadvantaged communities.
1
EP
Keywords: cooking energy, fuel switching, LPG, slum, India
AC C
Corresponding Author: (Present Address) D92, Second Floor, Shaheen Bagh, Jamia Nagar, New Delhi 110 025 India Tel: +91 9968100608 E-mail:
[email protected]
1
ACCEPTED MANUSCRIPT
1. Introduction
M AN U
SC
RI PT
Modern fuels – kerosene and liquefied petroleum gas (LPG) – are the major share (over 65%) of cooking energy in urban1 India. However, a large number of households, mostly disadvantaged, such as low-income and slum2 dwellers, still use traditional fuels – firewood, dung, crop residue, coal or charcoal. In comparison to modern fuels, traditional fuels adversely affect local indoor environment, which often result in poor health outcomes, particularly among women and children (de Koning et al., 1985; Holdren et al., 2000; Jerneck and Olsson, 2013; McMichael et al., 2000; Smith, 1993). The use of traditional fuels is thermally inefficient and diverts substantial fuel carbon to product of incomplete combustion, whose energy commitment per meal is high (Smith et al., 2000). It also depletes forest covers, as much of the firewood tends to come from illegal logging (Gautam and Herat, 2000), accentuating climate change and biodiversity loss. Fuel switching has several benefits, thus an example of ‘development with cobenefits’ (Puppim de Oliveira, 2013; Smith and Haigler, 2008).
EP
TE D
The household fuel choices depend upon three factors. The first is availability and access to fuels, the second is affordability, as determined by household income, and the third is policy options available, such as prices, subsidies, and taxes. Some of these factors are aggravated with spatial location of the household. In rural India, for instance, freely available traditional fuels and poor access to modern fuels discourage the use of LPG or kerosene as primary cooking fuel, even among high-income households. Similarly, slum dwellers largely use traditional fuels because of inadequate access to modern fuels and poor socio-economic conditions. Thus, specific conditions influence domestic energy patterns. Previous studies focus on urban and rural households, but silent about slum dwellers. This was the motivation to investigate similarity (or difference) in domestic energy patterns and switching factors between slum and non-slum households in urban areas. The large size of slum dwellers, about 65 million (17.2% of urban Indian), provide further impetus for such analyses (NBO, 2010). Thus, evidence-based interventions in slums can contribute to the environmental sustainability and improve quality of life, an important target under the Millennium Development Goals (Moreno, 2003).
AC C
The major contribution of this paper is to bring forth debates of cooking fuels switching of slum dwellers in the literature and practice. This study analyses the intra-urban differences among the different urban residents (slum and non-slum). Earlier studies on fuel shifting have focused mostly in urban or rural areas, or a comparison of both (Farsi et al., 2007; Heltberg, 2004; Rao and Reddy, 2007) , but hardly any study has attempted to understand the large intra urban differences among different urban areas, which is important for policy design for transition to modern fuels in all urban households. Since the socio-economic characteristics vary significantly between different urban inhabitants (Table 1), identical policy interventions may not work effectively for all. Another contribution of this study is establishing linkage between energy security and multidimensional aspects of urban deprivations. This study recognizes multiple cooking fuels used by the household and not limit to primary (or secondary) cooking fuel as 2
ACCEPTED MANUSCRIPT
often considered in earlier studies (Rao and Reddy, 2007; van der Kroon et al., 2013). Thus, we grouped household fuel use in three groups – only traditional, only modern and mixed – and analyze accordingly.
M AN U
SC
RI PT
The results reveal cooking fuels consumption patterns are consistent with the energy ladder theory in slum and non-slum households. In addition to income, prices of modern fuels, access to urban amenities (electricity and water supply), and educational attainment are the major determinants for opting modern fuels. Available policy intervention to promote uptake of modern fuels are the subsidies in fuels prices, where every Indian household is entitled for 12 subsidized cooking gas cylinders (14.2 Kg) in a year, while ‘eligible’ household is also entitled for a fixed amount of kerosene from the public distribution system (PDS). These subsidies are mistargeted and financially unsustainable. Departing from existing subsidy regime, this study recommends two major policy interventions: employing differential subsidies on modern fuels based on multidimensional aspects of poverty; and partially diverting existing subsidy to improve physical and social capitals, which would also improvised slums. The rest of the paper organizes as following. Section 2 presents brief review of the literature and section 3 presents data and descriptive statistics. Section 4 discusses estimations methods of fuel choice model. Section 5 presents estimation of results and section 6 concludes with policy implications of the results.
TE D
2. Previous studies
AC C
EP
Previous studies related to cooking fuels have focused on the patterns of fuel use, determinants of fuel switching, and demand assessment for fuels in developing economies. Recently studies also focuses on exploring efficient stoves and converting available solid fuels into biomass gasification (Purohit, 2009). These studies are descriptive and empirical in nature, where, descriptive studies primarily present patterns of fuels use (Alam et al., 1998; Reddy and Srinivas, 2009), while empirical studies estimate fuels demand and choice, including their drivers, using econometric methods (Farsi et al., 2007; Ouedraogo, 2006; Reddy, 1995). Results reveal income is one of the determinants of modern fuels choice and therefore supports the energy ladder theory. In other words, households adopt better quality of fuels with income improvement. Besides income, other factors influence fuels choices, such as fuels characteristics, socioeconomic characteristics, demographic patterns, urbanization level and access to electricity (Heltberg, 2004; Heltberg, 2005; Hosier and Dowd, 1987; Masera et al., 2000). Therefore, energy use patterns described by energy ladder theory are more complex than is being perceived. Moreover, studies reveal household uses multiple cooking fuels and socio-cultural and taste preferences matter in fuels choices. Among other interesting points, fixed costs associated with the modern fuels, such as connection cost of LPG act as a barrier (Israel, 2002). Several studies analyse fuel use patterns and fuel switching factors, but application of discrete choice framework is limited. Reddy (1995) using a series of binomial logit model on 500 3
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
households survey from Bangalore concludes that energy-carrier follows energy ladder. Besides income, family size and occupation of the head of the household also matters for fuel switching. Another four sets of studies, using the 1999- 2000 household expenditure survey from the National Sample Survey (NSS), have employed discrete choice framework for distinct research inquiries (Farsi et al., 2007; Heltberg, 2004; Rao and Reddy, 2007; UNDP/ESMAP, 2003). While evaluating effectiveness of prevalent subsidy on the modern fuels in facilitating fuel switching, UNDP/ESMAP (2003) reveals existing subsidies are skewed towards middle- and high-income groups, therefore, unable to achieve social objectives and fiscal sustainability. Heltberg (2004) employing multinomial logit model on comparable rural and urban household surveys from 8 countries, including India, reveals significant positive correlation between electrification and uptake of modern cooking fuels. Farsi et al. (2007) model fuel choices in urban household for three main cooking fuels – firewood, kerosene and LPG – using ordered discrete choice framework and concludes besides income and price, several socio-demographic factors, such as education and gender of the head of the households, are important determinants of fuel choice. Finally, using the same dataset Rao and Reddy (2007) estimates energy carrier choice decisions in urban and rural areas using multinomial logit model and concludes that income, household size, educational status of the head of the household, occupation of the household members and locational characteristics influence decision to choose a particular energy carrier.
TE D
Gupta and Köhlin (2006), using a 500 households survey from Kolkata, predicted choices of the cooking fuel. Using household stated ranking, study reveals that fuel attributes (price, availability and ease of use) and household socio-economic characteristics (consumption expenditure, education, age of the household head, religion and caste) determine fuel choices. To enhance the use of modern fuels, this study sought for improving availability of LPG and awareness of ill effects of traditional fuels, instead of subsidizing modern fuels.
AC C
EP
Though dual characteristics of urbanization in developing economies, including India, are well documented (Ahmad et al., 2013; Roy, 2005), but most of the studies on the fuel switching consider urban as a composite entity. Existing large differences in socio-economic characteristics and access to services between slums and non-slums demand disaggregated analysis. The next section briefly describes dual characteristics of urban India. 3. Data source and descriptive statistics This study uses micro data derived from the Indian Human Development Survey, conducted by the National Council of Applied Economic Research (NCAER), New Delhi and University of Maryland, covering entire urban India during the period November 2004 to October 2005 (Desai et al., 2010). The sample survey consists of 13,126 urban households, using stratified random sampling, from 971 urban blocks. Empirical analysis uses only 5,763 non-slum and 669 slum households because of missing values of some variables, mostly fuel prices. This could be corrected by replacing with the average prices of the fuels from immediate larger geographical 4
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
unit (say urban block or district), but refrained from doing so since the sample size was adequately large for the econometric analysis. The survey collects information about the use of six cooking fuels – firewood, dung, crop residue, coal or charcoal, kerosene and LPG – and their purpose of use – mainly cooking, mainly lighting, mainly heating and combination of these. The data set also includes information about source of traditional fuels collected or purchased and if collected then time taken in collection. The primary reason for using this data set over others, such as NSS consumer expenditure surveys, is disaggregation of data set by intra-urban areas (slum and non-slum), and adequate information about cooking fuels, including multiple uses of fuels. The households categorize into three groups by the use of cooking fuels: only traditional, only modern and mixed fuels. As mentioned earlier, the traditional fuels include firewood, dung, crop residue coal or charcoal while modern fuels include kerosene and LPG. The dataset has also information either the household come from the six largest metro cities (Delhi, Mumbai, Kolkata, Chennai, Hyderabad, and Bengaluru) or not, which is used as dummy variable to capture city-size effect.
TE D
Fig. 1 illustrates the relationship between income and fuel choices in non-slum and slum households. As income increases, probability of choosing modern fuels rises whereas traditional fuels and mixed fuels drop, in non-slum and slum households. This suggests that a moderateincome (third quintile) is likely to be associated with the fuel switching from traditional/mixed to modern in non-slum households, while traditional/mixed fuels are prevalent in slum households. A high-income (fourth and fifth quintiles) is likely to be associated with the fuel switching from traditional/mixed fuels to modern fuels in non-slum households, but less evident in slum households. Insert Figure 1 here
AC C
EP
The share of traditional fuels use in non-slum and slum households are 13% and 27% resepectively, while modern fuels use is significant in non-slum (65%) and slum households (41%). Relatively, higher share of slum dwellers use mixed fuels (32%) than non-slums (23%). As expected, slum dwellers have poor socio-economic conditions than non-slums, in almost all parameters (Table 1). Compared to non-slums, slum household have low-income, large household size, low education level and high concentration of disadvantaged communities (scheduled castes, scheduled tribes and other backward classes). Slum areas have cheaper firewood but coastly LPG and kerosene, partly on access and availability grounds. Slum dwellers also face difficulty in getting required documents to access susidized modern fuels. Insert Table 1 here 4. Model and Estimation methods Literature review reveals that most of the studies presume energy ladder theory for fuel choices and use of multinomial logit model. This model similar to theory of traditional innovation for 5
ACCEPTED MANUSCRIPT
any technology adaptation has distinct levels of ladder – learning, growth, saturation and declination (Heltberg, 2004; Hosier and Dowd, 1987; Masera et al., 2000). In early stage, households use traditional fuels, then mixed and finally modern fuels (Fig.2). Typically the user move step by step and leap frog is unlikely (Heltberg, 2004; Kowsari and Zerriffi, 2011).
RI PT
Insert Figure 2 here
M AN U
SC
The multinomial logit model (MNLM) is efficient estimate of the binary logit (Long and Freese, 2006), therefore, we employ MNLM, which assumes simultaneously estimating binary logit of all comparison among the alternatives – (a) traditional fuels, (b) mixed fuels, and (c) modern fuels. The goal of this model is to find the best fitting model to describe relationship between outcome variables and traditional/modern fuels use on base of mixed fuels. The MNLM is the extension of the logistic model where dependent variable (Y) is with ‘J’ nominal outcomes (in our case J = 3). There is need to calculate ‘J - 1’ equations one for each category with the reference category (b = mixed fuels). These J equations solutions predict the probabilities as shown in Eq.1. for m =1 to J
(1)
5. Results and discussions
TE D
Element of the vector βj measure effects of change in the corresponding variables on the probability that consumer i will choose fuel alternative j relative to the probability to choose mixed fuels.
AC C
EP
Table 2 presents results of the multinomial logit regressions of fuel switching in non-slum and slum households. The results estimate predictors of fuel switching to only traditional and only modern fuels on the base of mixed fuels. The models also include odd ratio to add interpretations. All models are without sampling weights and slum model do not use state dummy due to small sample size. To enhance our understanding about the substitution patterns amongst three forms of cooking fuels, Table 3 presents the marginal effects of the variables at the sample means.The values in this table can be interpreted as the effects of a one-unit change in a given independent variable (or a switch in the case of dummy variables) on the probability of choosing a particular fuel. In addition, Table 4 presents the marginal effects for price of LPG for different income groups of the population.
5.1 Effects of fuels prices
6
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
The results reveal that LPG price have significant and negative effect on non-slum households, suggesting that increasing price can result to the use of inferior quality of fuels. While kerosene price negatively affects fuel choice decision for both slum and non-slum household and thereby vindicates the hypothesis of the present study. It means increasing the price of kerosene increases probability to use modern fuels and reduces traditional fuels use. This ambiguity might occur due to not acknowledging the sources of kerosene – market or public distribution system (PDS) – in the analysis. In fact, relatively low-income households access kerosene from the PDS at cheaper price, while affluent households buy from the market. Recent migrants (usually urban poor) do not have access to kerosene from the PDS because of not possessing ration card. In addition, households are most likely to switch to LPG within the modern fuels category and even reduce energy consumption in the aftermath of increasing kerosene price. On average, less than one-fifth of the households primarily use kerosene3 for cooking purpose. Farsi et al. (2007) also reported similar ambiguous effect of kerosene price on its use. Increase in the price of firewood also induces fuel switch from traditional to mixed fuels in slum households. Insert Table 2 here
TE D
The marginal effects result suggest that a 10% decrease in LPG price will increase the average share of modern fuels users by 3.8%, while decreasing the share of traditional fuels by 1% in non-slum households (Table 3). As discussed earlier, the effect of kerosene is ambiguous. A 10% increase in kerosene price will increase the average share of modern fuel users (kerosene is the part of it) by 1.7% in non-slum and 2.2% in slum households Insert Table 3 here
AC C
EP
Differences in the marginal effects of LPG prices are evident by different income groups in nonslum and slum households. When facing higher LPG prices, the high- and middle-income households of non-slum households are more likely to shift to mixed fuels, while as expected, the low-income group substitute to traditional fuels. While the high-income slum households are more likely to substitute to mixed fuels. In non-slum, a 10% increase in LPG price can decrease the share of modern fuels users by 3.5% and increase traditional fuels share by 3.1% in lowincome group. While, in middle- and high-income groups the share of modern fuels users decrease by 2% and 1.4% respectively and increase mixed fuel share by 1.3% and 1.4% resepectively. In slum households, a 10% increase in LPG prices have negligible effects on the low-income households, while the high-income households decrease the share of modern fuels users by 9% and increase mixed fuels shares by 5.5%. Insert Table 4 here 5.2 Effects of households income As discussed in descriptive statistics of section 3, the share of modern fuels users increases by income quintiles both in non-slum and slum households, but at a slower pace in slum dwellers. 7
ACCEPTED MANUSCRIPT
RI PT
The econometric analyses also reveal that household income have a positive and significant effects on the use of modern fuels, providing other variables constant. The marginal effects at the sample mean reveal that 10% increase in income lead to increase the share of modern fuels users by 1% both in non-slum and slum households. As expected, households living below the poverty line (BPL dummy) have less probability to use modern fuels in comparison to the households living above the poverty line (APL) in non-slum areas. In resultatnt of switching household from BPL to APL, the share of modern fuels users increased by 21%. However, the variable – BPL dummy – has negative but not significant association with the use of modern fuels in slum dwellers.
SC
5.3 Effects of access to urban amenities
TE D
M AN U
Access to urban amenities – supply water and electricity – have positive and significant association with the use of modern fuels both in non-slum and slum households. The influence of access to electricity are greater than the access to supply water. Heltberg (2004) also shows such positive relation , however, due to lack of time series data this study also conclusively cannot prove that household choices rather than absolute supply factor derive such results. Electrified non-slum households are on an average 43% more likely to use modern fuels and about 30% less likely to use mixed fuels in comparison to non-electrified households. The effect of access to electricity is also significantly large in slum dwellers, where electrified households are on an average 33% more likely to use modern fuels and almost all shift come from traditional fuels users. In a similar way, households with access to supply water are on an average about 14.5% in non-slum and 23.5% in slum households likely to use modern fuels than households without access to supply water. Beside location, these results suggest that expanding urban amenities have stronger bearing with the use of modern fuels and such association in slum households are notable given their meager access to electricity.
EP
5.4 Effects of sociodemographic and other variables
AC C
As expected, education attainment have a positive and significant effects on modern fuels use. Notably, the impact of female education attainment is more than male. It means the education gap4 between genders hinder the uptake of modern fuels. With education improvemnet amongst female and male members of the households, the probability to use modern fuels increases. A 10% increase in education attainment level, for instance, increases the average share of modern fuels by 20 – 25% in non-slum and 23 – 33% in slum households. Thus, educational initiatives may have larger impact in slums. The lower value is for male and the upper value is for female education attainment. This finding underscore importance of education in general and particularly among female members. Disadvantaged communities have significantly negative effect on the use of modern fuels. Households living in metro cities (Mumbai, Delhi, Kolkata, Chennai, Banglore and Hyderabad) have high probability to use modern fuels in comparison to those living in small/medium urban canters, ceteris paribus, perhaps due to access to fuels. Nonslum model includes state dummy and a few of them have significant coefficients, suggesting 8
ACCEPTED MANUSCRIPT
RI PT
that there are differences in the choice behaviour of the households living in different provinces.5 The size of the household has negative and significant effect on choosing modern fuels in nonslum and slum households. It is understandable that the large families may divert their resources to fulfill other basic needs, as well as they can overcome some of the problems associated with the use/collection of traditional fuels.
6. Conclusion and policy implications
TE D
M AN U
SC
This paper presents the results of discrete choice models on fuel choices, and patterns of cooking fuels, using a nationally representative dataset of urban slum and non-slum households. The empirical analyses determine the responsiveness of fuel choices on fuels prices, income, access to urban amenities, and socio-demographic characteristics, among others. This research differs from previous studies in the following aspects: focuses on intra-urban differences (slum and nonslum), and considers all forms of cooking fuels, beyound primary cooking fuels. The results suggest that the observed patterns are consistent with the stylized “energy ladder” theory in slum and non-slum households. The use of modern fuels increases with income improvement. Initially, the effect of income seems to be less pervasive on switching to modern fuels in slums, but after controlling other variables, the effect is similar to non-slums (Table 3). Additionally, upholding the findings of recent studies, fuels prices, access to urban amenities and sociodemographic factors are important determinants for fuel choices (Farsi et al., 2007; Ouedraogo, 2006).
AC C
EP
From a policy point of view, a subsidization of LPG, promotion of education particularly among female members, expanding access to electricity and supply water, and overall economic development could be effective instruments to promote the use of modern fuels. As increasing price of the LPG would significantly affect disadvantaged communities, including low-income non-slum and high-income slum households. Thus, present flat subsidy on LPG is neither appropriate nor financially viable for uptake of modern fuels use. Given the high fiscal costs associated with LPG subsidy, we suggest differential price subsidies based on the multidimensional aspects of urban poverty as suggested by the S.R. Hashim Committee (Planning Commission, 2012). The committee considers multiple categories of vulnerabilities (residential, occupational and social) to identify and classify poor households. Some of the criteria used by the committee, such as access to electricity and socially disadvantaged communities, are also barriers to the use of modern fuels. Therefore, adequate subsidization of modern fuels for these households, rather than whole population, would increase the probability to use modern fuels and reduce the fiscal costs. The differential subsidies are also in accordance with the growing acceptance of the multidimensional understanding of poverty. Moreover, Aadhaar, a nationwide unique identification of each resident addresses the challenge associated with the identification of such households (see http://uidai.gov.in/).
9
ACCEPTED MANUSCRIPT
RI PT
The physical and social capitals, notably access to electricity and supply water, and education, have greater influence than the prices of the modern fuels. Therefore, to achieve social objective of fuel switching, it would be reasonable to divert existing fuel subsidy partially on improving social and physical capitals, with greater focus on disadvantaged communities. These interventions would helpful in achieving fuel switching and inclusive growth, which would address some of the severe problems of the slum dwellers.
M AN U
SC
Moreover, still a large proportion of urban households use only traditional fuels or in combination to modern fuels, particularly amongst disadvantaged communities, such as lowincome and slum dwellers. Therefore, other strategies should be laid down to reduce ill effects of the use of traditional fuels, such as provision of ventilated cooking spaces, efficient less polluting biomass cooks stoves, reforestation and afforestation program to produce renewable firewood. In the long term, renewable firewood can be an alternative to produce gasified biomass that could be even more sustainable alternative to the modern cooking fossil fuels. Awareness about cobenefits of fuels use, particularly health, would be also helpful for desirable fuels switching.
Acknowledgements
Notes 1
TE D
The first author is grateful to the Japan Society for the Promotion of Science (JSPS) for support through the JSPS-UNU Postdoctoral fellowship. The article is the sole responsibility to the authors and does not express the views of the institutions that they are affiliated with.
AC C
EP
Urban areas are defined as: (a) all statutory places with a municipality, corporation, cantonment board or notified town area committee, etc. (b) a place satisfying the following three criteria simultaneously: (i) a minimum population of 5,000; (ii) at least 75 per cent of male working population engaged in non-agricultural pursuits; and (iii) a density of population of at least 400 per sq. km. 2 Slum areas constitute of: (a) all specified areas in a town or city notified as ‘slum’ by public agencies under any Act including a ‘Slum Act’. (b) all areas recognized as ‘slum’ by public agencies including Housing and Slum Boards, which may have not been formally notified as slum under any act; and (c) a compact area of at least 300 population or about 60-70 households of poorly built congested tenements, in unhygienic environment usually with inadequate infrastructure and lacking in proper sanitary and drinking water facilities. All households not living in slum areas are non-slum. 3 About 15% non-slum and 23% slum households use Kerosene mainly for cooking, rest use for other purposes e.g., lighting, heating or combination of these. 4 For instance, census of India (2011) shows gap in literacy rate was 16.68 percentage points (male 82.14% and female 65.46%. Of course, in term of highest education attainment adult female lacks in comparison to adult male (Table4). 5 Result has not been reported but can be requested from the author.
10
ACCEPTED MANUSCRIPT
References
AC C
EP
TE D
M AN U
SC
RI PT
Ahmad, S., Balaban, O., Doll, C.N.H., Dreyfus, M., 2013. Delhi revisited. Cities 31, 641-653. Alam, M., Sathaye, J., Barnes, D., 1998. Urban household energy use in India: efficiency and policy implications. Energy policy 26, 885-891. de Koning, H.W., Smith, K., Last, J., 1985. Biomass fuel combustion and health. Bulletin of the World Health Organization 63, 11. Desai, S., Vanneman, R., National Council of Applied Economic Research, N.D., 2010. India Human Development Survey (IHDS), 2005. Inter-university Consortium for Political and Social Research (ICPSR) [distributor]. Farsi, M., Filippini, M., Pachauri, S., 2007. Fuel choices in urban Indian households. Environment and Development Economics 12, 757-774. Gautam, R., Herat, S., 2000. Environmental issues in Nepal and solving them using the cleaner production approach. Journal of Cleaner Production 8, 225-232. Gupta, G., Köhlin, G., 2006. Preferences for domestic fuel: analysis with socio-economic factors and rankings in Kolkata, India. Ecological Economics 57, 107-121. Heltberg, R., 2004. Fuel switching: evidence from eight developing countries. Energy Economics 26, 869887. Heltberg, R., 2005. Factors determining household fuel choice in Guatemala. Environment and Development Economics 10, 337-361. Holdren, J.P., Smith, K.R., Kjellstrom, T., Streets, D., Wang, X., Fischer, S., 2000. Energy, the environment and health. New York: United Nations Development Programme. Hosier, R.H., Dowd, J., 1987. Household fuel choice in Zimbabwe:: An empirical test of the energy ladder hypothesis. Resources and Energy 9, 347-361. Israel, D., 2002. Fuel Choice in Developing Countries: Evidence from Bolivia. Economic Development and Cultural Change 50, 865-890. Jerneck, A., Olsson, L., 2013. A smoke-free kitchen: initiating community based co-production for cleaner cooking and cuts in carbon emissions. Journal of Cleaner Production 60, 208-215. Kowsari, R., Zerriffi, H., 2011. Three dimensional energy profile: A conceptual framework for assessing household energy use. Energy Policy 39, 7505-7517. Long, J.S., Freese, J., 2006. Regression models for categorical dependent variables using Stata. Stata press. Masera, O.R., Saatkamp, B.D., Kammen, D.M., 2000. From linear fuel switching to multiple cooking strategies: a critique and alternative to the energy ladder model. World development 28, 20832103. McMichael, A., Confalonieri, U., Githeko, A., Haines, A., Kovats, R., Martens, P., Patz, J., Sasaki, A., Woodward, A., 2000. Human health. Special Report on Methodological and Technological Issues in Technology Transfer: A Special Report of IPCC Working Group III, 329-347. Moreno, E.L., 2003. Slums of the world: the face of urban poverty in the new millennium?: monitoring the millennium development goal, target 11--world-wide slum dweller estimation. Un-Habitat. NBO, 2010. Report of the Committee on Slum Statistics/Census. Ministry of Housing and Urban Poverty Alleviation, Governemnet of India, New Delhi. Ouedraogo, B., 2006. Household energy preferences for cooking in urban Ouagadougou, Burkina Faso. Energy Policy 34, 3787-3795. Planning Commission, 2012. Report of the Expert Group to Recommended the Detailed Methodology for the Identification of Families Living Below Poverty Line in the Urban Areas Planning Commission of India, Delhi.
11
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Puppim de Oliveira, J.A., 2013. Learning how to align climate, environmental and development objectives in cities: lessons from the implementation of climate co-benefits initiatives in urban Asia. Journal of Cleaner Production 58, 7-14. Purohit, P., 2009. Economic potential of biomass gasification projects under clean development mechanism in India. Journal of Cleaner Production 17, 181-193. Rao, N., Reddy, B.S., 2007. Variations in energy use by Indian households: An analysis of micro level data. Energy 32, 143-153. Reddy, B.S., 1995. A multilogit model for fuel shifts in the domestic sector. Energy 20, 929-936. Reddy, B.S., Srinivas, T., 2009. Energy use in Indian household sector – An actor-oriented approach. Energy 34, 992-1002. Roy, A., 2005. Urban informality: toward an epistemology of planning. Journal of the American Planning Association 71, 147-158. Smith, K.R., 1993. Fuel combustion, air pollution exposure, and health: the situation in developing countries. Annual Review of Energy and the Environment 18, 529-566. Smith, K.R., Haigler, E., 2008. Co-benefits of climate mitigation and health protection in energy systems: scoping methods. Annu. Rev. Public Health 29, 11-25. Smith, K.R., Uma, R., Kishore, V., Zhang, J., Joshi, V., Khalil, M., 2000. Greenhouse implications of household stoves: an analysis for India. Annual Review of Energy and the Environment 25, 741763. UNDP/ESMAP, 2003. Access of the Poor to Clean Household Fuels in India. World Bank, p. 94. van der Kroon, B., Brouwer, R., van Beukering, P.J.H., 2013. The energy ladder: Theoretical myth or empirical truth? Results from a meta-analysis. Renewable and Sustainable Energy Reviews 20, 504-513.
12
ACCEPTED MANUSCRIPT
Table 1: Characteristics of non-slum and slum households in urban India, 2005
Description
Only traditional fuel [DV] Mixed fuel use [DV]
firewood, dung, crop residue and coal/charcoal mixed fuels (traditional and modern)
Only modern fuel [DV]
Kerosene and/or LPG
Price fire-wood
Rs./kg
Price LPG
Rs./kg
Price kerosene-market
Rs./liter
Income
BPL dummy
total household income with all means in Rs. (‘000) number of family member in the household number of meals including breakfast taken by the household the highest school attainment among adult men (year of schooling) the highest school attainment among adult female (year of schooling) household live below poverty line
Piped water dummy
dwelling unit has piped water supply
Electricity dummy
dwelling unit has electricity supply
hed-adult male hed-adult female
TE D
avmeals/day
Scheduled Caste households
Scheduled Tribes households
Social class - OBCs
Other Backward Classes households
Traditional fuel - collected
traditional fuel collected from field
Traditional fuel - purchase
traditional fuel purchased from market
Location-metro
household located in six metros
EP
Social class - SC Social class - ST
Pooled (n=14543) frequency(%) / mean 1946 (13.54%) 3312 (20.05%) 9108 (63.34%)
194.33***
1.87
-4.22***
21.83
3.37***
4.14
4.64
4.62
-8.98***
78.89
48.72
77.39
-7.95***
4.86
5.35
4.89
5.94***
2.80
2.93
2.92
-4.94***
9.42
6.69
9.28
-14.87***
7.31
4.08
7.15
-16.03***
0.17
0.37
0.18
14.08***
SC
1.73
21.99
0.79
0.84
0.79
3.30***
0.95
0.92
0.95
-3.78***
0.15
0.30
0.16
10.26***
0.03
0.06
0.03
4.42***
0.37
0.43
0.37
3.24***
0.27
0.27
0.27
-0.03
0.06
0.07
0.06
0.56
0.23
0.09
0.22
-8.69***
Notes: Categorical dependent variable [DV] is frequency; other figures are the mean values. The statistic is t-value in case of mean and chi-square value in case of frequency. The statistic is associated with testing the null hypothesis that the relevant non-slum and slum mean values were equal. ***: p-value<0.01, **: p-value<0.05, *: p-value<0.1
AC C
Statistic
21.82
M AN U
Household size
1.88
Slum (n=725) frequency(%) / mean 197 (27.48%) 228 (31.80%) 292 (40.73%)
RI PT
Variables
Non-slum (n= 13818) frequency(%) / mean 1749 (12.81%) 3084 (22.59%) 8816 (64.59%)
ACCEPTED MANUSCRIPT
Table 2: Multinomial logit of fuel switching in non-slum and slum households in urban India, 2005 non-slum Variable Coef.
odd ratio
slum modern
Coef.
traditional odd ratio 0.974 0.202 2.048 1.540 0.270 1.021 1.097 1.119 0.423 1.863 5.502
Coef.
odd ratio
modern Coef.
RI PT
traditional
odd ratio 1.395 1.561 2.034 1.525 0.347 0.907 1.083 1.126 0.707 3.409 7.203
AC C
EP
TE D
M AN U
SC
Price firewood (log) -0.171** 0.843 -0.027 -0.401 0.670 0.333 Price LPG (log) 1.887** 6.599 -1.601*** 0.971 2.640 0.445 Price kerosene (log) -0.678** 0.508 0.717*** -0.718** 0.488 0.710** Income (log) -0.290*** 0.748 0.432*** -0.077 0.926 0.422** HH size (log) 0.326** 1.385 -1.308*** 0.248 1.282 -1.059*** Meals per day 0.017 1.017 0.021 -0.302* 0.739 -0.098 Highest edu male -0.026* 0.974 0.092*** -0.073** 0.930 0.080*** Highest edu female -0.072*** 0.931 0.112*** -0.035 0.965 0.119*** BPL dummy 0.359*** 1.432 -0.861*** 0.36 1.433 -0.347 Piped water dummy -0.230* 0.795 0.622*** -0.08 0.923 1.226*** Electricity dummy -0.772*** 0.462 1.705*** -0.966** 0.381 1.975** Social gr. – SC (ref: 0.355 0.527*** 1.694 -0.584*** 0.557 -0.694* 0.499 -1.035*** others) Social gr. – ST (ref: 0.262 0.374 1.454 -0.549*** 0.577 -0.834 0.434 -1.338** others) Social gr. – OBCs (ref: 0.595 0.507*** 1.661 -0.313*** 0.731 -0.55 0.577 -0.519 others) traditional fuel 0.070 0.214* 1.239 -3.165*** 0.042 -0.112 0.894 -2.654*** purchased 1.199 Metro dummy -1.213*** 0.297 1.178*** 3.247 -1.927*** 0.146 0.182 1.395 -1.623 1.136 0.670 -7.213 Constant -1.674 State dummies used Y N Observations 5,763 669 Pseudo - R2 0.428 0.305 Notes: Base category: mixed fuels; State dummy is suppressed in non-slum and not used in slum; *** p<0.01, ** p<0.05, * p<0.1
ACCEPTED MANUSCRIPT
Table 3: Marginal effects at the sample mean non-slum mixed 0.006
traditional -0.100
Price firewood (log)
modern 0.108
Price LPG (log)
0.108
0.273
-0.381
0.140
-0.161
Price kerosene (log)
-0.043
-0.124
0.166
-0.188
-0.030
0.218
Income (log)
-0.021
-0.076
0.097
-0.048
-0.053
0.101
HH size (log)
0.046
0.234
-0.280
0.132
0.127
-0.258
Meals per day
0.000
-0.004
0.004
-0.047
0.044
0.003
-0.004
0.024
RI PT
modern -0.001
slum mixed -0.008
traditional -0.005
-0.003
-0.016
0.020
Highest edu female
-0.005
-0.020
0.025
-0.016
-0.014
0.030
BPL dummy
0.045
0.162
-0.207
0.096
0.010
-0.106
Piped water dummy
-0.028
-0.116
0.145
-0.100
-0.135
0.235
0.045
Social gr. – ST (ref: others)
0.037
Social gr. – OBCs (ref: others)
0.029
Traditional fuel - purchased
0.100
Metro dummy
-0.036
M AN U
-0.128
Social gr. – SC (ref: others)
SC
Highest edu male
Electricity dummy
-0.020
0.021
-0.301
0.429
-0.338
0.004
0.334
0.103
-0.149
-0.049
0.218
-0.169
0.100
-0.137
-0.067
0.268
-0.201
0.052
-0.081
-0.057
0.129
-0.072
0.563
-0.663
0.166
0.341
-0.507
-0.162
0.199
-0.227
0.083
0.144
AC C
EP
TE D
Notes: For dummy variables the effects are obtained from probability differences.
ACCEPTED MANUSCRIPT
Table 4: Marginal price effects of log (LPG) at the sample mean by income category
low 33% high 33% middle 33%
slum
low 33%
mixed 0.143 (11.9) 0.131 (23.2) 0.041 (32.4) 0.555 (26.0) -0.060 (31.3) 0.127 (37.8)
modern -0.143* (85.8) -0.204* (66.6) -0.351*** (41.8) -0.924** (58.6) -0.053 (41.4) -0.000 * (23.2)
RI PT
middle 33%
non-slum
traditional 0.000 (2.3) 0.073 (10.2) 0.311* (25.7) 0.368 (15.4) 0.114 (27.3) -0.127 (39.0)
SC
Income group high 33%
AC C
EP
TE D
M AN U
Notes: parenthesis values are share of fuels use by income group; *** p<0.01, ** p<0.05, * p<0.1; The average income of slum households in lower, middle and higher groups were 75%, 62% and 61% of the non-slum households respectively.
100 traditional
mixed
modern
RI PT
90 80 70 60 50
SC
40 30 20 10 0 1
2
3
4
M AN U
Percentage of households using the fuels mainly for cooking
ACCEPTED MANUSCRIPT
5
1
2
3
4
5
non-slum slum Quintile of household monthly income in non-slum and slum
Figure 1. Use of cooking fuels by income in non-slum and slum households in urban India, 2005
AC C
EP
TE D
Note: The average income by quintiles between non-slum and slum households vary as following (in ‘000 INR). Nonslum: 15.8; 34.1; 55.6; 88.2; and 201.9, and slum: 12.8; 23.7; 33.7; 51.8; and 121.8.
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Figure 2. Conceptual model of household energy ladder, 2012
ACCEPTED MANUSCRIPT
Highlights
RI PT
SC M AN U TE D EP
• •
Drivers of fuel switching were estimated in non-slum and slum households. The changes in fuel use patterns were consistent with the energy ladder theory. Fuel prices, urban amenities, and educational levels were found to be drivers for changing fuel usage. Improvement in physical and social capitals enhance modern fuels uptake. Differential fuel subsidies based on socio-economic characteristics were proposed.
AC C
• • •