Journal Pre-proof Household-level Analysis of Electricity Consumption on Welfare and Environment in Cambodia: Empirical Evidence and Policy Implications
Phoumin Han, Fukunari Kimura, Suwin Sandu PII:
S0264-9993(19)31473-7
DOI:
https://doi.org/10.1016/j.econmod.2019.11.025
Reference:
ECMODE 5083
To appear in:
Economic Modelling
Received Date:
16 September 2019
Accepted Date:
24 November 2019
Please cite this article as: Phoumin Han, Fukunari Kimura, Suwin Sandu, Household-level Analysis of Electricity Consumption on Welfare and Environment in Cambodia: Empirical Evidence and Policy Implications, Economic Modelling (2019), https://doi.org/10.1016/j.econmod.2019.11.025
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. 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. © 2019 Published by Elsevier.
Journal Pre-proof
Household-level Analysis of Electricity Consumption on Welfare and Environment in Cambodia: Empirical Evidence and Policy Implications Han Phoumin1, Fukunari Kimura2, and Suwin Sandu3 Corresponding author: Han Phoumin, Ph.D Affiliation: Economic Research Institute for ASEAN and East Asia (ERIA) based in Jakarta, Indonesia; Contact Address: Sentral Senayan II, 5th &6th Floor, Jl. Asia Afrika No.8., Jakarta Pusat 10270, Indonesia. Email address:
[email protected]; or
[email protected] Email address of co-author:
[email protected]
Energy Economist with Economic Research Institute for ASEAN and East Asia (ERIA) based in Jakarta, Indonesia; he can be contacted at [
[email protected]] 2 Professor of Economics at Keio University, and Chief of Economist with Economic Research Institute for ASEAN and East Asia (ERIA) based in Jakarta, Indonesia; he can be contacted at [
[email protected]] 3 Faculty of Engineering and Information Technology, University of Technology Sydney, Australia; he can be contacted at [
[email protected]] 1
Journal Pre-proof
Household-level Analysis of the Impacts of Electricity Consumption on Welfare and the Environment in Cambodia: Empirical Evidence and Policy Implications Phoumin Han,1 Fukunari Kimura,2 and Suwin Sandu3
Abstract Cambodia’s biomass consumption is the most dominant energy source at residential sector, and its use is mainly for cooking and heating which could affect health due to indoor air pollution. The biomass is mainly sourced from wood cutting and forest-encroachment that could impact the environment due to reduction of forest at considerable scale. By using the data 2015 of Cambodia Socio-Economic Survey, the study investigates the impacts of electricity consumption on household welfare, such as earnings and the school performance of children in the households, and further to investigate its impacts on the environment. The study found that household’s access to electricity with ability to spend on electricity consumption contributes to the positive household welfare effects and environment via a reduction of biomass consumption, and the more household spends on biomass, the more they are prone to sickness of lung problem. The study also confirmed the important role of human capital formation for the positive impact on the welfare and the environment. These findings lead to policy implications that would improve affordable access to electricity to ensure that all households can use electricity for their basic needs and productivity, and also to reduce the negative effects on environment.
Keywords: Electricity consumption, welfare, and environment JEL Classification: Q4 I00 O13
Phoumin Han is an energy economist at the Economic Research Institute for ASEAN and East Asia (ERIA); he can be contacted at
[email protected]. 2 Fukunari Kimura is professor of economics at Keio University and chief economist at the ERIA; he can be contacted at
[email protected]. 3 Suwin Sandu is a lecturer at the University of Technology, Sydney; he can be contacted at
[email protected]. 1
1
Journal Pre-proof
1.
Introduction
Cambodia has made enormous progress in providing electricity coverage – improving people’s access to electricity from 49.7% in 2016 to almost 80% in 2019 (Ministry of Mines and Energy, 2019). However, energy access remains low in rural areas and some remote areas have no electricity at all (Han and Kimura, 2019). The residential sector consumes much more energy than the commercial and industry sectors – about 73% of the total energy consumption of the three sectors (residential, commercial, and industry) in 2015. As shown in Figure 1, biomass was the main source of energy (about 87%) supplying the residential sector in 2015, followed by electricity at about 7% and oil products at 6%. Cambodian households use a variety of energy sources to meet their needs because of the high electricity price, e.g. some poor households use electricity for lighting but use biomass for cooking and heating (Han and Kimura, 2019). Electricity tariffs vary according to the user, i.e. in 2019, the electricity tariff is $0.185 per kilowatt-hour (kWh) for households with electricity consumption greater than 201 kWh/month, and $0.173/kWh for the commercial sector (EAC, 2019). Cambodia’s electricity tariff is the highest in Southeast Asia and has already threatened the competitiveness of production, as electricity is one of the major inputs for producing goods and services in the economy. The high production cost of goods at the current electricity tariff in textiles and footwear is offset by cheap labour and access to the European market under the European Union’s Generalised Scheme of Preferences, known as Everything But Arms, as Cambodia is a least developed country member of the World Trade Organization (European Commission, 2019). However, the Generalised Scheme of Preferences or Everything But Arms will be phased out in November 2019, and Cambodia will face huge challenges if its energy costs remain high.
2
Journal Pre-proof
Figure 1: Historical and Projected Energy Demand for the Residential Sector with Energy Sources 5.0
Biomass
4.5
Electricity
Oil Products
4.0 3.5 3.0 MTOE
2.5 2.0 1.5 1.0 0.5 0.0 BAU EEF 2015
APS BAU EEF 2020
APS BAU EEF 2025
APS BAU EEF 2030
APS
APS = alternative policy scenario, BAU = business as usual, EEF = energy efficiency framework, MTOE = million tons of oil equivalent. Source: Ministry of Mines and Energy (2019).
Access to grid electricity does not tell the whole story of how households use electricity. In many developing countries, insufficient consumption of electricity in households remains an important issue because electricity consumption is primarily a function of income and the price of electricity (Khraief et al., 2018), although some research has found that the causality of electricity consumption and income run both ways (Bridge, Adhikari, and Fontenla, 2016). Romero-Jordán, del Río, and Peñasco (2016) analysed the deep economic crisis and sharp rises in electricity prices during 2006–2012, and found that the increases in retail electricity prices had detrimental welfare effects, especially on the lower-income segment of the population. Nevertheless, household income seems to be the main determinant of electricity consumption in developing countries (Han and Kimura, 2019). Dong and Hao (2018), using provincial panel data for 1996–2013, confirmed the significant negative impact of urban–rural income disparities on provincial per capita electricity consumption. Hasan and Mozumder (2017) examined how energy use at the household level varies with income in Bangladesh, and found that households increase energy expenditure less proportionally as income grows, up to a threshold of household income. Other household characteristics also help explain electricity consumption. Sánchez-Sellero and Sánchez-Sellero (2019), using a stepwise regression analysis to discover the determinants of household electricity consumption and expenditure, confirmed that the surface area of a house, the number of people in a household, and electronic devices such as heating and hot water systems, are the major determinants of household electricity consumption. Similarly, Sharma, Han, and Sharma (2019) used primary data on various socio-economic variables 3
Journal Pre-proof
collected from 1,000 households in Mumbai to analyse the determinants of energy consumption, and found that an increase in the dwelling size of low- and very low-income group households increases electricity expenses. The educational levels of households also seem to influence the selection of energy types. Acharya and Marhold (2019) analysed the energy selection behaviour of Nepalese households using the annual household survey, and found that households with lower educational levels tend to select low-quality fuels such as firewood and kerosene, whereas ownership of information and communication technology devices lead to the selection of more modern and cleaner fuels. Households’ access to electricity is believed to have many health benefits. Tri Sambodo and Novandra (2019) studied the state of energy poverty and its impacts on welfare in Indonesia, and found that access to electricity and modern cooking fuels reduced the rate of malnutrition. Han and Kimura (2019) examined Cambodia’s energy poverty and its effects on social wellbeing, and found that poor households are strongly associated with the type of fuel used and low consumption of unaffordable clean energy, which has negative impacts on the health and well-being of the households. This study is inspired by households’ combined use of electricity with biomass consumption, even though they are connected to an electrical grid. Higher use of biomass in households has long-term effects on health, especially for young children and elderly persons. Furthermore, a higher share of biomass consumption in the residential sector creates more forest and shrub wood destruction, leading to an unsustainable environment. Since the use of electricity use is an important input for household productivity, either directly or indirectly, higher electricity consumption in households may induce productivity and income generation. Therefore, this study uses the national survey data of the Cambodia Socio-Economic Survey (CSES) 2015 (Government of Cambodia, 2015) to investigate the effects of household electricity consumption on welfare and the environment. The aspects of welfare considered in this study are household income, health effects such as respiratory diseases, and children’s schooling. Environmental impacts will be captured through the reduction of wood/biomass consumption as households increase electricity consumption. The findings will contribute to empirical evidence and theories, and will shape policy to address energy access and energy affordability in developing countries. The paper proceeds as follows. Section 2 describes gross domestic product (GDP), electricity demand growth, and tariff differences in Cambodia. Section 3 presents the data set, variables, and multicollinearity test. Chapter 4 presents the empirical approach and estimations. Chapter 5 discusses the results, and chapter 6 contains the conclusions and policy implications.
4
Journal Pre-proof
2.
GDP Growth, Electricity Demand, and Tariff Differences in Cambodia
Cambodia has achieved high economic growth in the past 6 years, averaging around 7% per year during 2014–2019 (Asian Development Bank, 2019). In 2019, the Asian Development Bank provided the latest economic data for Cambodia compared with other countries in Southeast Asia. The data show that Cambodia’s GDP growth rate is the highest (7%) in the Association of Southeast Asian Nations (ASEAN), followed by Viet Nam (6.8%), Myanmar (6.6%), the Lao People’s Democratic Republic (6.5%), the Philippines (6.4%), Indonesia (5.2%), Malaysia (4.5%), Thailand (3.9%), Singapore (2.6%), and Brunei Darussalam (1.0%). Empirical studies have shown that electricity demand is strongly explained by per capita GDP or income, and the demand is elastic in the long run to both income and price (Jamil and Ahmad, 2011; and Han and Kimura, 2014). Cambodia’s fast and stable economic growth is accompanied by strong growth in demand for electricity. Figure 2 shows that Cambodia’s electricity demand increased from 0.2 terawatt-hours (TWh) in 1995 to 4.4 TWh in 2015, and is projected to grow to 38.3 TWh in 2040.
Figure 2: Cambodia’s Historical and Future Electricity Demand by Energy Source
20.00 18.00
Coal
Oil
Natural gas
Hydro
Others
TeraWatt-hour (TWh)
16.00 14.00 12.00 10.00 8.00 6.00 4.00 2.00 0.00 0.20 0.45 4.40 10.61 1995 2000 2015 2020 Source: Authors’ calculation from Kimura and Han (2019).
14.03 2025
19.05 2030
26.57 2035
38.20 2040
Another key driver for the strong growth in electricity demand is population growth. Cambodia experienced 2.8% annual population growth during 1995–2015 and is expected to have a 1.5% annual growth rate from 2015 to 2040 (Kimura and Han, 2019). Rapid economic growth brings energy infrastructure development, including power grid extension and investment. By 2018, about 80% of households had access to grid electricity because of such extensions – allowing the government to exceed its goal of 70% of households 5
Journal Pre-proof
having access to electricity well in advance of the 2030 target date (Ministry of Mines and Energy, 2019). Nonetheless, electricity appears to place a large burden on households, given the high electricity tariff and the difference in tariffs between urban and rural areas. Cambodia’s electricity tariffs are complicated and vary between rural and urban areas; service providers; and residential, commercial, and industry customers – depending on whether electricity is purchased from independent power producers, own generation, or neighbouring countries. The Electricity Authority of Cambodia (EDC) is the main service provider, followed by 23 other service providers throughout the country. The electricity tariff ranges from $0.168/kWh for large consumers with medium-sized voltage meters to $0.238/kWh for small consumers at low voltage. For the residential sector, the electricity tariff is classified based on the consumer group or type, e.g. $0.21/kWh for rural areas and $0.15–$0.19/kWh for urban areas, varying from very small consumers whose electricity consumption is less than 10 kWh/month to large consumers whose electricity consumption is greater than 201 kWh/month (Ministry of Mines and Energy, 2019). The government plans to reduce the gap between rural and urban tariffs by introducing crosstariff consumer subsidies. This means that high electricity users will be charged more to subsidise the rural electricity tariff. By the 2020 plan (Ministry of Mines and Energy, 2019), the gap in electricity tariffs between rural and urban areas for medium voltage (22 kilovolts) will be reduced substantially – the rural tariff will be $0.162/kWh, while the urban tariff will be $0.129/kWh (Ministry of Mines and Energy, 2019). The high tariff gap between rural and urban areas is seen as a major obstacle to economic development, as most rural people are less wealthy than urban people but are subject to a higher electricity tariff than urban dwellers. This situation precludes affordable electricity consumption in rural areas, and implies many welfare impacts and loss of economic opportunities from using electricity because of the high cost of this energy. Furthermore, even urban dwellers cannot enjoy significant electricity access because the tariff remains expensive compared with neighbouring countries. Thus, both rural and urban areas face high energy costs and only wealthy households can enjoy sufficient electricity consumption.
3.
Econometric Approach
Two main issues are delineated above. The first is whether electricity access and electricity consumption affect households’ earning opportunities. The second is whether electricity access and consumption have impacts on households’ welfare and the environment. The econometric approaches are formulated based on the above questions. The study also reviewed past studies and traditional approaches supporting the estimation, and fine-tuned the econometric model to test the above questions.
6
Journal Pre-proof
3.1.
Impacts of Electricity Access and Consumption on Household Income
Since the beginning of the industrial revolution, literacy and knowledge have become increasingly valuable relative to basic manual skills. This increasing value has led to wage premiums for educated workers. Not surprisingly, an educated workforce is the dominant factor in explaining differences in regional growth and prosperity. As a result, economists have extensively researched the importance of education in determining individual differences in wages and regional differences in economic growth. The works on investment in human capital by Becker (1962, 1975) emphasised education and training – the most important investments. Of course, formal education is not the only way to invest in human capital. Workers also learn and are trained outside schools, especially on the job. A number of studies have used Mincer’s human capital earnings function because this model is the most commonly employed method in labour economics. Mincer (1974) used a simple regression model with a linear schooling term and a low-order polynomial in potential experience. However, our data capture only the binary variable of the household head’s completion of schooling (e.g. primary, secondary, high school, or university), so the empirical specification of the household’s earnings or income will include this information. The variable of interest is the electricity consumption per capita, since this study wishes to test whether electricity consumption has a positive impact on the earnings or income of a household. Other variables of household and community characteristics, such as household expenditures, assets, and rural/urban locations, are included as control variables. Thus, the model specification to test the question is simply the multivariate regression model as follows:
𝐼𝑛𝑐𝑜𝑚𝑒𝑖 = 𝛼0 + 𝛽1𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑖𝑡𝑦𝑖 + 𝛽2𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑖𝑡𝑦2𝑖 + 𝛽3𝐸𝑑𝑢𝑖 + 𝛽4𝐸𝑥𝑝𝑒𝑛𝑑𝑖 + 𝛽5𝑋𝑖 + 𝑈𝑖1
(eq.1)
𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑡𝑦𝑖 = 𝛾0 + 𝛿1𝐼𝑛𝑐𝑜𝑚𝑒𝑖 + 𝛿2𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝐷𝑒𝑣𝑖𝑐𝑒𝑠𝑖 + 𝑈𝑖2
(eq.2)
Where in equation (1), variable 𝐼𝑛𝑐𝑜𝑚𝑒𝑖 is the household income, 𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑖𝑡𝑦𝑖 is the per capita household electricity consumption, and 𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑖𝑡𝑦^2𝑖 is the square of the per capita household electricity consumption. The inclusion of 𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑖𝑡𝑦^2𝑖 will capture the nature of the trade-off relationship between the household income and the electricity use. Non-linearity of electricity consumption as a control variable will try to measure some saturated level of electricity consumption or any wasteful behaviour of households that use electricity carelessly regardless of income. The variable 𝐸𝑑𝑢𝑖 is the education level of the household head. The variable 𝐸𝑥𝑝𝑒𝑛𝑑𝑖 is the total household expenditure, while 𝑋𝑖 refers to other household and community characteristics. 7
Journal Pre-proof
Where in equation (2), the explanatory variable 𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝐷𝑒𝑣𝑖𝑐𝑒𝑠𝑖represents the instrumental variables such as television (TV), number of cell phone, and computer, which are correlated with the variable 𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑖𝑡𝑦𝑖, but they are not correlated with error term𝑈𝑖1. Since the explanatory variable 𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑖𝑡𝑦𝑖and its square is expected to be endogenous with the error term 𝑈𝑖1, the structural equation (1) will suffer biased coefficients estimate if we use the Ordinary Least Square (OLS) method. Thus, we also test this endogeneity and we can reject the hypothesis that 𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑖𝑡𝑦𝑖is exogenous. To deal with this endogeneity, the structural equation (1) will be estimated by using the instrumental variables which are the variables of electricity devices such as television (TV), number of cell phone, and computer. Thus, the estimated equations (1) & (2) are the simultaneous equation, or the Two Stage Least Square (2SLS) estimations. In equation (1) and equation (2), the 𝑈𝑖1 and 𝑈𝑖2are zero-mean error terms, and the correlations between 𝑈𝑖1and the elements of 𝑈𝑖2are presumably nonzero.
3.2.
Impacts of Electricity Access and Consumption on Households’ Welfare and the Environment
In the literature, welfare refers to a type of government support that ensures people can meet their basic human needs (e.g. food and shelter) and other services such as healthcare, education, and vocational training (Department for Works and Pensions, 2019). This study wants to test the question regarding household electricity consumption and its impact on the welfare of the household and the environment. Welfare, as defined in this study, refers to the food and non-food consumption, and health of the household. Environment refers to the household’s choice of using electricity instead of biomass. Wood consumption contributes to forest encroachment and will destroy forest reserve if the trend of biomass consumption continues to grow in Cambodia. An economic theory of welfare, known as the Bergson–Samuelson social welfare function, is based on Bergson (1938, 1954) and Samuelson (1956). The theory uses the utility function and takes the following form: 𝑊 = 𝑊(𝑈1,𝑈2,..,𝑈𝐻); where society welfare is denoted as W, which is the function of the utilities of its constituent members, 𝑈𝐻; and h = 1, 2, ..,H, where H denotes the number of households in the society. The theory is simply the so-called Pareto-optimal points of the utility function, which represents the social desirability. Rawls (1971) developed a social welfare function known as the Rawlsian social welfare function, which takes the form of 𝑊 = 𝑚𝑖𝑛(𝑈1,𝑈2,..,𝑈𝐻) to maximise the utility of society’s least happy member, which is also known as Leontief-type social indifference curves.
8
Journal Pre-proof
Adopting the concept of the utility function, the household’s welfare function is believed to be affected enormously by the household’s condition and situation of accessing electricity and the amount of electricity consumption and other household and community characteristics. This household welfare function could be expressed as follows: 𝑊𝑒𝑙𝑓𝑎𝑟𝑒 = 𝑊(𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑖𝑡𝑦𝑖,𝐼𝑛𝑐𝑜𝑚𝑒𝑖, 𝐴𝑠𝑠𝑒𝑡𝑠𝑖..,𝑈𝑖), where i = 1, 2,.., N of the households. Since parents’ education forms an important part of explaining the welfare of children in the household, it could also be included in the utility function. From this utility function, the study formulates the empirical model to capture the impact of electricity consumption on the household’s welfare, and the empirical model is extended to capture the impact on the environment through the reduction of the household’s biomass consumption. The empirical models are as follows: 𝑊𝑒𝑙𝑓𝑎𝑟𝑒𝑖 = 𝑣0 + 𝑏1𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑖𝑡𝑦𝑖 + 𝑏2𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑖𝑡𝑦^2𝑖 + 𝑏3𝐼𝑛𝑐𝑜𝑚𝑒𝑖 + ᴓ2𝑋𝑖 + 𝑈𝑖2
(Eq.3)
𝐸𝑛𝑣𝑖𝑟𝑜𝑛𝑚𝑒𝑛𝑡𝑖 = 𝑐0 + 𝑑1𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑖𝑡𝑦𝑖 + 𝑑2𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑖𝑡𝑦^2𝑖 + 𝑑3𝐼𝑛𝑐𝑜𝑚𝑒𝑖 + 𝑘2𝑋𝑖 + 𝑈𝑖3 (Eq.4)
Where dependent variable 𝑊𝑒𝑙𝑓𝑎𝑟𝑒𝑖 represents the household’s welfare such as food and nonfood consumption, respiratory illnesses, and children’s school performance. The dependent variable 𝐸𝑛𝑣𝑖𝑟𝑜𝑛𝑚𝑒𝑛𝑡𝑖 is the biomass consumption of the household, where a reduction in biomass consumption is expected to have a positive impact on the environment. The explanatory variables 𝐼𝑛𝑐𝑜𝑚𝑒𝑖 and 𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑖𝑡𝑦𝑖 are the income and per capita electricity consumption of the household. The variable 𝑋𝑖 is the household and community characteristics. The 𝑋𝑖 variable also includes the household’s education, as this information on the household head – especially the education of the mother and father – has been shown to have an impact on the human capital accumulation of the child. See, for instance, Ray (2000); Deb and Rosati (2004); Blunch, Sudharshan, and Goyal (2002); Heady (2003); and Khanam (2003). These studies confirmed the positive link between parents’ education and the likelihood of a child attending school. The household size is also included as an explanatory variable because the dependent variable is per capita income or expenditure, which covers biases for household size. The outcome variable of children’s school performance is the Index of Schooling Attainment Relative to Age (SAGE) adopted from Han (2008), in which the index took the percentage derived from [SAGE=((Year of schooling)/(Age-Age of school entry)*100))]. The higher the index, the better the schooling outcomes for children. Other factors, especially social infrastructure, play a significant role in increasing welfare. However, differences in social infrastructure are only characterised by geographical areas such as rural and urban.
4.
Data Set, Variables, and Multicollinearity Test 9
Journal Pre-proof
The sample of households with access to grid and off-grid electricity is drawn from the CSES 2015 data set. The National Institute of Statistics of Cambodia carried out the CSES 2015 from January to December 2015 on a sample of 3,840 households. The CSES 2015 is a survey of households and their members on housing conditions, education, economic activities, production and income, household level and structure of consumption, health, victimisation, vulnerability, energy source and consumption, and others. The CSES database at the National Institute of Statistics is open to external researchers for research and analysis. Since this study focuses on households connected to grid and off-grid electricity, the sample used in the regression analysis is restricted to this purpose. In the sample, 2,790 households had access to grid electricity, and the remaining sample 1,050 households were considered not connected to grid electricity. The variables of interest in this study are described below. The variable for per capita household income in United States (US) dollars is derived from the total household income, divided by the size of the household, and transformed into a logarithm to use in the regression. The unit used in the per capita household income is the US dollar. The per capita household expenditure in US dollars variable is derived from a similar procedure to the per capita household income variable. The lung problem variable represents respiratory problems and takes the value of 1 any of the household members is reported to have had lung problems in the past 6 months and the value of zero otherwise. The SAGE variable is the Index of Schooling Attainment Relative to Age – a higher index represents higher school performance of the child. Biomass, a binary variable, denotes the biomass reported to have been used by households for cooking and heating for the last 7 days. Other explanatory variables are electricity per capita (kWh/year); human capital of the household head (primary school complete, secondary school complete, and university); household assets (televisions, cars, cell phones, and computers); gender of the household head; and other community characteristics (rural and urban). These variables are shown in Table 1. Table 1: Description of Variables and Statistics Variable Log per capita income of the household ($/year) Log per capita expenditure of the household ($/year) Lung problem = 1 if any household member reported to have respiratory problem in the past 6 months Index of school attainment relative to age (%) Biomass = 1 if household is reported to have used the biomass for cooking and heating in the past 7 days Household’s expenditure on biomass ($/year) Per capita electricity consumption of the household (kWh/year) Household head with primary education complete Household head with lower secondary education complete
Obs. 2,790 2,791
Mean 6.90 6.95
Std. Dev. .830 .536
Min. -1.09 5.18
Max. 11.48 9.76
2,791
.048
.214
0
1
2,063 2,791
77.5 .52
27.7 .49
5.5 0
100 1
2,791 2,791
4.44 298.56
6.41 317.09
0 17.16
112.5 7,146
2,791 2,791
.206 .135
.40 .34
0 0
1 1
10
Journal Pre-proof
Household head with university complete Number of televisions in the household Number of cell phones in the household Number of computers in the household Number of cars in the household Male household head Rural
2,791 2,791 2,791 2,791 2,791 2,791 2,791
.05 .98 2.07 .20 .09 .74 .46
.23 .54 1.64 .69 .30 .43 .49
0 0 0 0 0 0 0
1 5 11 20 2 1 1
kWh = kilowatt-hour, Max. = maximum = Min. = minimum, Obs. = observations, Std. Dev. = standard deviation. Source: Authors’ calculation.
The study performed the variance inflation factor (VIF) to detect multicollinearity in the regression analysis – to check the correlations between predictors or independent variables in the model. The VIF = 1/(1-Ri2 ) estimates how much the variance of the regression coefficient is inflated because of multicollinearity. The results in Table 2 show that all variables used in the regression are good and acceptable.
Table 2: Variance Inflation Factor Variable Log per capita electricity consumption of the household Log per capita electricity consumption of the household square Biomass Household expenditure on biomass Rural Log per capita expenditure of the household Number of televisions Number of cell phones Number of cars Household head with university complete Household head with lower secondary school complete Number of computers Household head with primary school complete School attainment relative to age Male gender household head Lung problem Mean VIF
VIF 3.94 3.89 2.17 1.73 1.52 1.51 1.43 1.32 1.28 1.25 1.15 1.15 1.13 1.10 1.06 1.02 1.73
1/VIF 0.202314 0.257086 0.461117 0.577315 0.657920 0.660153 0.700566 0.760392 0.780967 0.796966 0.870129 0.873297 0.886601 0.907584 0.942872 0.977939
VIF = variance inflation factor. Source: Authors’ calculation.
5.
Results and Discussions
Table 3 shows the instrumental regression coefficient estimates of the household income, in which the main hypothesis is that accessing and using electricity provides opportunities for income-generating activities, either directly or indirectly. However, since the household income or earnings have been studied intensively in the past, and the main determinant of household income is human capital formation such as education and experience, the information is included as a regressor. The coefficient estimate of per capita electricity consumption and its square are statistically significant, with a positive sign on the level form and a negative sign on 11
Journal Pre-proof
the non-linear form. This means that households use electricity for productivity either directly or indirectly. Taking the partial derivative with respect to the per capita electricity consumption variable, we found that households’ electricity consumption contributes to household income. This clearly confirmed our hypothesis that electricity plays a significant role in household income. Nevertheless, we also ran the regression for the off-grid sample of households in remote rural areas that are not connected to grid electricity but where some households use electricity from small power generation distributors, which have limited operational hours and high prices. People still use kerosene for lighting in rural Cambodia. Electricity from private providers, where available, is very expensive – as high as $1/kWh (Han, 2019) – and only operates for about 2–3 hours in the evening (generally 6:00–9:00 pm). This study aims to assess if such small consumption of electricity contributes to household income. The coefficient estimate of per capita electricity consumption for the off-grid sample size is insignificant. This confirms that a very small amount of electricity consumption does not contribute to household productivity or income. Other variables such as human capital (household head education) are positive and statistically significant. These results confirm the positive relationship between human capital formation and household income and earnings. The results support the past literature and theory regarding the return on education from human capital investment. The rural coefficient is statistically insignificant for the on-grid and off-grid sample. This indicates that rural areas are disadvantaged by the insignificant amount of electricity consumption, which prevents households from using electricity as an input for production and earnings. Table 3: Instrumental Variable Regression of Per Capita Household Income Function Dependent variable: Log per capita income of the household ($/year)
Instrumental variable coefficient estimates Sample size: On-grid Sample size: Off-grid
Per capita electricity consumption of the household (kWh/year) Per capita electricity consumption of the household square (kWh/year) Human capital variables Household head with primary school complete Household head with lower secondary school complete Household head with university complete Others Male gender Rural
12
.0064067*** (.001955) -2.41e-06** (9.03e-07)
.0310513 (.0199255) -.0000397 (.0000259)
.0264488 (.1014148) .1186135 (.1257531) .0746635 (.2475741)
-.1698227 (.3040442) -.2670724 (.5178732) -1.029529 (2.531632)
-.0383131 (.1049449) .1703579 (.1670607)
.339268 (.3944233) -.1402185 (.5393605)
Journal Pre-proof
Constant
5.39445*** (.4683718)
3.461925 (2.362237)
Tests of endogeneity for the instrumental variables regression (sample size: On-grid) Ho: variables are exogenous Durbin (score) chi2 (2) = 51.0563 (p = 0.0000) Wu-Hausman F(2,2779)= 25.9015 (p = 0.0000) Tests of endogeneity for the instrumental variables regression (sample size: Off-grid) Ho: variables are exogenous Durbin (score) chi2 (2) = 38.6024 (p = 0.0000) Wu-Hausman F(2,1036) = 19.8295 (p = 0.0000) kWh = kilowatt-hour. Notes: Number of observations = 2,790; Wald chi2(8) = 77.98; Prob > chi2 =0.0000; R-squared = 0.2209; Root MSE = 1.9126; ***, **, and * represent the significant level at 99%, 95%, and 90% respectively. Source: Authors’ calculation.
The study also questioned whether households’ welfare improves if they obtain access to electricity. The welfare effects defined in this study are household expenditure, lung problems caused by exposure to household biomass consumption, and the school performance of children in the household. There is a widely held assumption that sufficient electricity consumption leads to better welfare, as it improves household expenditure, health (less exposure to indoor air pollution), and children’s school performance (additional time reading and doing homework with artificial light). Electricity is one of the items in the basket of household consumption. Increasing electricity consumption will improve household welfare, as households benefit from electricity access and consumption, thus it will increase household expenditure. Therefore, household income plays a major role in determining overall household expenditure. However, households prioritise items in the household consumption basket when faced with budget constraints. Thus, the electricity consumption explains total household expenditure. Table 4 shows the instrumental regression coefficient estimates of log per capita household expenditure. The coefficient of per capita electricity consumption is statistically significant and positive. This means that electricity consumption explains the increase in household expenditure, and household welfare will increase as a result. Other variables such as human capital formation and assets help explain the overall increase in household expenditure. Thus, these results confirm the role of electricity in improving welfare; and substantiate the theory of the economic welfare of the population, which states that the increasing the bundle of consumption is necessary for supporting welfare and quality of life. The insignificant sign of the rural variable means that household expenditure in rural areas relies on the basket of consumption mainly comprises food whereas rural households easily obtain it from subsistence farming.
13
Journal Pre-proof
Table 4: Instrumental Variable Regression of Per Capita Household’s Expenditure Function Dependent variable: Log per capita expenditure of the household ($/year) Per capita electricity consumption of the household (kWh/year) Per capita electricity consumption of the household square (kWh/year) Human capital variables Household head with primary school complete Household head with lower secondary school complete Household head with university complete Others Male gender Rural Constant
Instrumental variable coefficient estimates Sample size: On-grid Sample size: Off-grid .0033185*** (.0008272) -1.04e-06** (3.82e-07)
.0176917* (.0092048) -.0000198 (.000012)
.0845963* (.0428989) .1389118** (.0532) .1330321** (.044825)
-.0446763 (.1404571) -.0406468 (.2392381) -.6964597 (1.169519)
-.0106099 (.0443926) .1284842 (.0707073) 6.057759*** .1982111
.2604611 (.1822088) .1063248 (.2491644) 4.606413 (1.091265)
Tests of endogeneity for the instrumental variables regression (sample size: On-grid) Ho: variables are exogenous Durbin (score) chi2(2) = 28.7953 (p = 0.0000) Wu-Hausman F(2,2780) = 14.4904 (p = 0.0000) Tests of endogeneity for the instrumental variables regression (sample size: Off-grid) Ho: variables are exogenous Durbin (score) chi2(2) = 22.2947 (p = 0.0000) Wu-Hausman F(2,1036) = 11.2702 (p = 0.0000) kWh = kilowatt-hour. Notes: Number of observations = 2,791; Wald chi2(8) = 244.87; Prob > chi2 =0.0000, R-squared = 0.1209, Root MSE = .80908; ***, **, and * represent the significant level at 99%, 95%, and 90% respectively. Source: Authors’ calculation.
Other negative welfare effects, such as lung problems of household members, are seen as a consequence of long exposure to the use of biomass for cooking and heating. Thus, the study also questions the negative effects of using biomass and the positive effects of accessing and using electricity to improve the quality of life. It is assumed that households with enough electricity will forgo the use of biomass, although it is abundant and available (either free or inexpensive). However, the decision to consume enough electricity to meet a household’s basic needs depends on the cost of electricity and the household income. Therefore, households connected to grid electricity still use a combination of electricity and biomass. 14
Journal Pre-proof
Table 5 shows the coefficient estimates of the probit regression of lung problems. The per capita electricity consumption coefficient is statistically significant and has a negative sign, whereas the household expenditure on biomass coefficient is statistically significant and has a positive sign. These results mean that households using electricity tend to reduce the probability of their household members contracting lung problems such as respiratory diseases. However, households using biomass or combining biomass with electricity tend to have a high probability of their household members having respiratory diseases.
Table 5: Probit Regression Log-Likelihood Estimates of Lung Problems Dependent variable: Lung problem Per capita electricity consumption of the household (kWh/year) Household’s expenditure on biomass ($/year) Per capita income of the household ($/year) Human capital variables Household head with primary school complete Household head with lower secondary school complete Household head with university complete Assets Number of televisions
Multivariate coefficient estimate with robust standard error correction (Full sample size: On and off-grid) -.0007732** (.0003582) .0246296*** (.0063713) 8.81e-06 (.0000107) .0061261 (.108621) .0083729 (.1363246) -.1762507 (.3087391) .2009281* (.1051286) -.0749178** (.0321333) -.2340383* (.1190241) .1293716 (.1625067)
Number of cell phones Number of computers Number of cars Others Male gender
.2408797** (.1054343) .0457302 (.101968) -1.823643*** (.1621979)
Rural Constant
kWh = kilowatt-hour. Notes: Probit regression number of observations = 2,791; Wald chi2(12) = 65.72; Prob > chi2 = 0.0000; Log pseudolikelihood = -505.56249; Pseudo R2 = 0.0648; (***), (**), and (*) represent the significant level at 99%, 95%, and 90% respectively. Source: Authors’ calculation.
15
Journal Pre-proof
Access to electricity facilitates many household activities besides household chores. Children’s performance at school relies heavily on how much time they spend doing homework in the evening. Thus, electricity is expected to help increase children’s school performance. Table 6 shows the coefficient estimates of the multivariate regression of school attainment relative to age: SAGE. The coefficient estimates of both per capita electricity consumption and household expenditure on education are statistically significant and positive. These results mean that accessing electricity, together with household expenditure on children’s education, contribute to children’s improved school performance. However, although primary and secondary education are supposed to be free as part of Cambodia’s education for all policy, other expenses such as school materials and extra classes are enormous – requiring out-of-pocket payments to support schoolchildren. High informal fees such as paying for extra class hours, imposed on students by teachers to offset their low wages, also prevent children from attending school (Millar, 2018). Other important variables – human capital formation and access to television – are statistically significant and positive. These findings confirm the correlation between the human capital formation of parents or household heads and children’s education (Han, 2018).
Table 6: Regression Coefficient Estimates of ‘Schooling Attainment Relative to Age: SAGE’ Dependent variable: ‘SAGE: Schooling Attainment Relative to Age’
Per capita electricity consumption of the household (kWh/year) Per capita electricity consumption of the household square (kWh/year) Household expenditure on education ($/year) Human capital variables Household head with primary school complete Household head with lower secondary school complete Household head with university complete Assets Number of televisions
Multivariate coefficient estimate with robust standard error correction (Full sample size: On and off-grid) .0067299* (.0040884) -1.95e-06* (1.01e-06) .0087183*** (.0014865) 6.70222*** (1.298779) 6.701692*** (1.528147) 14.73078*** (2.58716) 7.364928*** (1.022149) -.5491767 (.3449017) -.6793233 (1.218807) -1.394914 (1.62578)
Number of cell phones Number of computers Number of cars
16
Journal Pre-proof
Dependent variable: ‘SAGE: Schooling Attainment Relative to Age’
Others Male gender
Multivariate coefficient estimate with robust standard error correction (Full sample size: On and off-grid) -.0205464 (1.241917) -1.434216 (1.270248) 63.47619*** (1.906415)
Rural Constant
kWh = kilowatt-hour. Notes: Number of observations = 2,838; F(12, 2825) = 28.88; Prob > F = 0.0000; R-squared = 0.1243; root MSE = 26.941; ***, **, and * represent the significant level at 99%, 95%, and 90% respectively. Source: Authors’ calculation.
The study also examined how accessing and using electricity could have impacts on the environment. As stated above, biomass represents 87% of the energy use in the residential sector, so significant amounts of wood need to be cut to meet that demand. This level of consumption of biomass would definitely affect the environment, as wood is natural carton stock. Households’ access to electricity and their ability to use electricity to replace biomass are considered positive for welfare and the environment. Table 7 shows log-likelihood estimates of using biomass. The coefficients of both per capita electricity consumption and per capita expenditure are statistically significant and have a negative sign. These results suggest that both accessing electricity and the ability of households to spend on electricity have a high probability of reducing biomass consumption. Thus, households’ shift from biomass to electricity will have a positive impact on the environment. Other important variables such as human capital formation and access to television have statistically significant coefficients and negative signs. This confirms the important role of human capital formation, such as the educational level of the household head and access to television, in households switching from biomass use. In other words, it has a high probability of reducing the biomass consumption of that household.
17
Journal Pre-proof
Table 7: Probit Regression Log-Likelihood Estimates of ‘Using Biomass’ Dependent variable: “Using biomass”
Per capita electricity consumption of the household (kWh/year)
Per capita expenditure of the household ($/year)
Multivariate coefficient estimate with robust standard error correction (Full sample size: On and off-grid) -.0005682** (.000199) -.0002874*** (.0000778)
Human capital variables Household head with primary school complete Household head with lower secondary school complete Household head with university complete Assets Number of televisions
-.2341802** (.069597) -.2314768** (.0860044) -.8067372*** (.1805802) -.2659907*** (.0744521) -.109954*** (.0191404) -.0918637 (.0575384) -.1742483 (.1194327)
Number of cell phones Number of computers Number of cars Others Male gender
.1441756** (.0637093) 1.152983*** (.0596349) .573053*** (.135103)
Rural Constant
Notes: Probit regression number of obs = 2,791; Wald chi2(11) = 823.02; Prob > chi2 = 0.0000; Log pseudolikelihood = -1321.7485; Pseudo R2 = 0.3158; (***), (**), and (*) represent the significant level at 99%, 95% and 90% respectively. Source: Authors’ calculation.
6.
Conclusions and Policy Implications
This study was carried out to highlight the importance of households’ access to and consumption of electricity, as they have many positive effects on households’ welfare and the environment – especially in Cambodia, where households rely on biomass consumption for cooking and heating. The study tested two main questions regarding the effects of electricity access and consumption on welfare and the environment. Using the CSES 2015 – the national survey data on households and its members with regard to households’ characteristics such as
18
Journal Pre-proof
income, expenditure, main activities, security, housing, energy, agriculture, and others – the study employed multivariate and probit regression models to analyse the questions. The findings suggest that households with electricity access and the ability to spend on electricity consumption contribute to household welfare – such as rising household incomes, increasing children’s school performance, reducing the risk of household members falling sick with respiratory problems, and potentially reducing the biomass consumption or shifting from biomass to clean electricity use. However, households which are connected to electricity but have low purchasing power to spend on electricity will still combine the use of biomass and electricity. The resulting impacts remain to be seen, as they depend on the intensity of the biomass use. The more a household spends on biomass, the more they are prone to respiratory illnesses. For people in remote or rural areas, the low levels of household electricity consumption do not have positive impacts on welfare. This is due to the high electricity cost in rural areas, as electricity is operated by small private distributors who sell it at very high prices. As a result, households in rural areas use electricity just 1–2 hours in the evening for dinners or special activities, which does not contribute significantly to households’ income-generating activities or any improvement in welfare. In some rural areas, households rely completely on kerosene and biomass for heating and cooking. The study confirmed the important role of human capital formation, such as the educational level of the household head and access to information such as media and television, in creating a positive impact on the welfare and the environment. The study has significant policy implications: o The current reliance on biomass consumption in the residential sector is 87%. Apart from its deleterious effects on human health, this high dependency has many negative environmental effects because of the forest encroachment undertaken to obtain the biomass. Thus, shifting away from biomass towards clean electricity is an urgent policy to electrify the entire population with access to clean energy. This study proved that access to electricity is not enough on its own, and it will not solve the problem of households switching away from biomass. The underlying root cause of households’ decisions to switch to electricity depends on their purchasing power – their ability to spend on electricity. These results suggest that poor households require policy attention and that any electricity policy subsidy will need to target poor households. Thus, data on poor households will need to be updated so that electricity subsidies are not wasteful. o Off-grid households are not connected to grid electricity. However, they can access privately generated electricity at rates five to 10 times higher than urban on-grid electricity rates. Households can only afford 1–2 hours of electricity consumption in the evening, putting the rural poor at a significant disadvantage as they face higher energy costs than better-off households in urban areas. Thus, electricity cross-subsidies need to be considered for rural and off-grid households. The government will also need to accelerate the construction of rural energy infrastructure such as the national grid or 19
Journal Pre-proof
mini-grid extensions to reach rural areas. This will require investment and appropriate policy support. o Another possibility is to electrify rural areas through an off-grid distributed electricity system from solar rooftop, micro-hydropower, and/or biomass power plants. However, the government would need to investigate the system cost of this energy and ensure that it is affordable for rural people. One possibility is for the government to procure a large number of solar panels and remove all tariffs related to solar, micro-hydropower, and biomass boiler and other related materials for small distributed energy systems. Alternatively, local solar photovoltaic set-ups could reduce the cost and make the system affordable in rural areas.
References: Acharya, B. and K. Marhold (2019), ‘Determinants of Household Energy Use and Fuel Switching Behaviour in Nepal’, Energy, 169, pp.1132–38. Asian Development Bank (2019), Cambodia: Economy. Economic Indicators for Cambodia. https://www.adb.org/countries/cambodia/economy (accessed 16 August 2019). Becker, G. S. (1962), ‘Investment in Human Capital: A Theoretical Analysis’, Journal of Political Economy, 70(5), pp.9–49. Becker, G. S. (1975), Human Capital: A Theoretical and Empirical Analysis, 2nd Ed. New York: Columbia University Press (for National Bureau of Economic Research). Bergson, A. (1938), ‘A Reformulation of Certain Aspects of Welfare Economics’, The Quarterly Journal of Economics, 52(2), pp.310–34. Bergson, A. (1954), ‘On the Concept of Social Welfare’, The Quarterly Journal of Economics, 68(2), pp.233–52. Blunch, N.-H., C. Sudharshan, and S. Goyal (2002), ‘Short- and Long-term Impacts of Economic Policies on Child Labor and Schooling in Ghana’, Social Protection Discussion Papers and Notes, No. 0212. Washington, DC: World Bank. Bridge, B.A., D. Adhikari, and M. Fontenla (2016), ‘Household-level Effects of Electricity on Income’, Energy Economics, 58, pp.222–8. Deb, P. and F. Rosati (2004), ‘Estimating the Effects of Fertility Decisions on Child Labor and Schooling’, A Joint Research Project of the ILO, World Bank and the United Nations Children’s Fund (UNICEF). Department for Works and Pensions (2019), About Us. www.gov.uk/government/organisations/department-for-work-pensions/about (accessed 19 August 2019). Dong, X.-Y. and Y. Hao (2018), ‘Would Income Inequality Affect Electricity Consumption? Evidence from China’, Energy, 142(C), pp.215–27. Electricity Authority of Cambodia [EAC] (2019), Retrieved from [https://www.eac.gov.kh/document/tariffdecidelist]. Accessed date 12 July 2019. 20
Journal Pre-proof
European
Commission
(2019),
Policy,
Countries
and
regions:
http://ec.europa.eu/trade/policy/countries-and-regions/countries/cambodia/
Cambodia. (accessed
17 July 2019). Government of Cambodia, National Institute of Statistics (2016), Cambodia Socio-Economic Survey, 2015. Phnom Penh. http://www.nis.gov.kh/nis/CSES/Final%20Report%20CSES%202015.pdf (accessed 7 June 2019). Han, P. (2015), ‘Energy Security and Sustained Growth: Analysis of the Energy Outlook and Savings Potential in the EAS Region’, Asia Pathways (The Blog of the Asian Development Bank Institute), 18 November. https://www.asiapathways-adbi.org/2015/11/energysecurity-and-sustained-growth-analysis-of-the-energy-outlook-and-savings-potential-inthe-eas-region/ (accessed 18 February 2019). Han, P. (2008), ‘Human Capital and Hours Worked of Children in Cambodia: Empirical Evidence and Policy Implications’, Asian Economic Journal, 22(1), pp.25–46.
Han, P. and S. Kimura (2014), ‘Analysis of Price Elasticity of Energy Demand in East Asia: Empirical Evidence and Policy Implications for ASEAN and East Asia’, ERIA Discussion Paper Series, No. 5, Jakarta: ERIA. Han, P. and F. Kimura (2019), ‘Cambodia’s Energy Poverty and its Effects on Social Wellbeing: Empirical Evidence and Policy Implications’, Energy Policy, 132, pp.283–89. Hasan, S.A. and P. Mozumder (2017), ‘Income and Energy Use in Bangladesh: A Household Level Analysis’, Energy Economics, 65(C), pp.115–26. Heady, C. (2003), ‘The Effect of Child Labor on Learning Achievement’, World Development, 31(2), pp. 385–98. International Energy Agency (2017a), Energy Access Outlook 2017: From Poverty to Prosperity – World Energy Outlook Special Report. Paris: IEA. International Energy Agency (2017b), Southeast Asia Energy Outlook 2017 – World Energy Outlook Special Report. Paris: IEA. Jamil, F. and E. Ahmad (2011), ‘Income and Price Elasticities of Electricity Demand: Aggregate and Sector-Wise Analyses’, Energy Policy, 39(9), pp.5519–27. Khanam, R. (2003), ‘Child Labor and School Attendance: Evidence from Bangladesh’, mimeo, University of Sydney. Khraief, N., M. Shahbaz, H. Mallick, and N. Loganathan (2018), ‘Estimation of Electricity Demand Function for Algeria: Revisit of Time Series Analysis’, Renewable and Sustainable Energy Reviews, 82(3), pp.4221–34. Kimura, S. and P. Han, eds. (2019), Energy Outlook and Energy Saving Potential in East Asia 2019. Jakarta: ERIA. Kutani, I., ed. (2013), Study on the Development of an Energy Security Index and an Assessment of Energy Security for East Asia Countries, ERIA Research Project Report 2012, No. 24. Jakarta: ERIA. Millar, P. (2018) ‘The Sobering Reality of Cambodia’s Free Education Drive’, Southeast Asia Globe, 1 October. https://southeastasiaglobe.com/the-sobering-reality-of-cambodiasfree-education-drive/ (accessed 20 June 2019). 21
Journal Pre-proof
Mincer, J.A. (1974) Schooling, Experience, and Earnings. New York: Columbia University Press (National Bureau of Economic Research). Ministry of Mines and Energy (2019), Cambodia Basic Energy Plan. ERIA publication. Otaka, Y. and P. Han, eds. (2016), Study on the Strategic Usage of Coal in the EAS Region: A Technical Potential Map and Update of the First-Year Study. Jakarta: ERIA. Rawls, J. (1971), A Theory of Justice as Fairness. Cambridge, MA: Harvard University Press. Ray, R. (2000), ‘Child Labor, Child Schooling, and Their Interaction with Adult Labor: Empirical Evidence for Peru and Pakistan’, The World Bank Economic Review, 14(2), pp.347–67. Romero-Jordán, D., P. del Río, and C. Peñasco (2016), ‘An Analysis of the Welfare and Distributive Implications of Factors Influencing Household Electricity Consumption’, Energy Policy, 88, pp.361–70. Samuelson, P.A. (1956), ‘Social Indifference Curves’, The Quarterly Journal of Economics, 70(1), pp.1–22. Sánchez-Sellero, M.-C. and P. Sánchez-Sellero (2019), ‘Variables Determining Total and Electrical Expenditure in Spanish Households’, Sustainable Cities and Society, 48, 101535. Sharma, S.V., P. Han, and V.K. Sharma (2019), ‘Socio-Economic Determinants of Energy Poverty Amongst Indian Households: A Case Study of Mumbai’, Energy Policy, 132, pp.1184–90. Tri Sambodo, M. and R. Novandra (2019), ‘The State of Energy Poverty in Indonesia and its Impact on Welfare’, Energy Policy, 123(132), pp.113–21. United Nations (2011), 2012 International Year of Sustainable Energy for All. http://www.un.org/en/events/sustainableenergyforall/ (accessed 19 February 2019). United Nations (2017), The Sustainable Development Goals Report 2017. New York: UN. World Bank (2016), Sustainable Development Goal on Energy (SDG7) and the World Bank Group. http://www.worldbank.org/en/topic/energy/brief/sustainable-developmentgoal-on-energy-sdg7-and-the-world-bank-group (accessed 21 February 2019). World Bank (2019), Access to electricity: % of population. https://data.worldbank.org/indicator/EG.ELC.ACCS.ZS (accessed 21 February 2019).
22
Journal Pre-proof
Highlights
o Household’s access to electricity with ability to spend on electricity consumption contributes to the positive household welfare effects. o The more household spends on biomass, the more they are prone to sickness of lung problem. o The study confirmed the important role of human capital formation for the positive impact on the welfare and the environment.
1