- Email: [email protected]
The impacts of energy insecurity on household welfare in Cambodia: Empirical evidence and policy implications
Economic Modelling xxx (xxxx) xxx
Contents lists available at ScienceDirect
Economic Modelling journal homepage: www.journals.elsevier.com/economic-modelling
The impacts of energy insecurity on household welfare in Cambodia: Empirical evidence and policy implications Han Phoumin a, *, Fukunari Kimura b a b
Energy Economist with Economic Research Institute for ASEAN and East Asia (ERIA) Based in Jakarta, Indonesia Keio University, Chief Economist with Economic Research Institute for ASEAN and East Asia (ERIA) Based in Jakarta, Indonesia
A R T I C L E I N F O
A B S T R A C T
JEL classification: Q48 C39 I39
This study investigates the impacts of energy insecurity on household welfare in Cambodia. The notion of energy insecurity is not well understood in the literature as well as in local contexts. This study defines household energy insecurity as the status quo derived from the interplay of inadequate and insufficient energy consumption that prevents households from meeting basic energy needs. The notion of energy insecurity can only be well understood by investigation in the local context as it varies from place to place. Households facing insufficient energy consumption may forgo many other opportunities. Once energy security has been defined in the Cambodian context, the study employs multiple regression models using the Cambodia Socio-Economic Survey Data (2015) to investigate the impacts of household energy insecurity. The study confirmed that energy insecurity has enormous negative impact on welfare of the households, with a further negative impact on the human capital formation of the children. The findings will lead to policy implications to improve household energy security, and thus impact economic, social, and environmental development.
Keywords: Energy insecurity Schooling and welfare
1. Introduction The concept of household’s energy security is the 1st step to be defined to get general understanding of its meaning and scope. In the literatures, the word “security” means “the state of being free from danger or threat”. Adopting this definition, the household’s energy security takes to mean the interplay of adequate and sufficient consumption of households to meet basic energy needs. In other word, the household is energy insecurity when that household is in the state of inadequate and insufficient energy consumption to meet basic energy needs. As energy uses, especially the electricity and liquefied Natural Gas (LPG) is centrally important for supporting the daily life of households in terms of cooking, lighting, reading, watching television, and other household income generating business activities, the inadequate and insufficient consumption of these energy will affect their quality of life at various degrees. Let alone, about 65 million people of Association of Southeast Asia Nations (ASEAN) has denied access to electricity and about 250 million people relies on biomass as a cooking fuel (IEA, 2017). Some countries in ASEAN such as Cambodia, Lao PDR and Myanmar, remote islands of Indonesia and Philippines, are facing challenges in providing electricity access in affordable price. Households having access to grid electricity is not the end of the problem as they face high
electricity tariff. For the better-off households the high electricity tariff may not impact much for their affordability of electricity consumption, but for the poor households, the electricity tariff may impact their affordability, leading to insufficient electricity consumption. Han and Kimura (2019) confirmed that access to electricity alone does not tell about adequate energy consumption. People with less wealth and income tend to use electricity only for important activities such as during dining and lighting for a few hours at night, leading to insufficient consumption. Energy insecurity exists not only in developing countries, but in developed countries as well; for instance, a study conducted by the US Energy Information Administration (EIA, 2018) found that about one-third of US households struggled to pay energy bills just to maintain adequate heating and cooling in their homes in 2015. According to the same report, about one in five households reported reducing or forgoing necessities such as food and medicine to pay an energy bill, and 14% reported receiving a disconnection notice for energy service. Households may also use less energy than they would prefer; 11% of households surveyed reported keeping their home at an unhealthy or unsafe temperature. In Europe, 50 million–125 million people in 2009 were struggling to meet their energy needs, including in some countries having undergone economic austerity reforms; for example, since the outbreak of the economic crisis of Greece in 2009, 64% of the country’s northern
* Corresponding author. E-mail addresses: [email protected] (H. Phoumin), [email protected] (F. Kimura). https://doi.org/10.1016/j.econmod.2019.09.024 Received 28 May 2019; Received in revised form 16 September 2019; Accepted 17 September 2019 Available online xxxx 0264-9993/© 2019 Published by Elsevier B.V.
Please cite this article as: Phoumin, H., Kimura, F., The impacts of energy insecurity on household welfare in Cambodia: Empirical evidence and policy implications, Economic Modelling, https://doi.org/10.1016/j.econmod.2019.09.024
H. Phoumin, F. Kimura
Economic Modelling xxx (xxxx) xxx
2.1. Supply security of energy
of import source countries, and strategic oil reserves. Most research studies on energy security refer to ‘securing of the amount of energy required for people’s lives, economic, social, and defense activities, among other purposes, at an affordable prices’ (Kutani, ed., 2013). In the East Asia Summit (EAS) context,1 the leaders adopted the Cebu Declaration in 2007 reaffirming the commitment to energy security by working closely together to reduce dependence on conventional fuels through intensified energy efficiency and conservation programmes and expansion of renewable energy systems and biofuel production/utilization (ASEAN, 2007). In recent years, ASEAN has striking towards diversifying its energy mix, and agreed on aspirational target to increase the share of renewable energy to 23% by 2025 in the primary energy mix (ACE, 2019). In recent development in the past few years, countries have been accelerating the installation of solar farm as well as the introduction of solar roof-top in many countries of ASEAN. Thanks for the speed of cost reduction of solar module cost and the gradual shift of know-how capacity to handle the solar and wind technologies into the ASEAN (The ASEAN Post, 2018). However, despite huge attempt and efforts by the leaders in ASEAN to ensure electricity access to all population, there remain challenges in some countries of ASEAN where electricity access remain low and electricity cost is high, and thus making both accessibility and affordability become core economic disadvantage as well as threatening the energy security of the household as these accessibility and affordability threatening daily life of people (Han and Kimura, 2019). From a broad view, the predicted primary and final energy demand from 2015 to 2040, which is expected to grow almost double in the EAS region during the forecasting periods, is posing an increasing threat to energy security (Han, 2015). Further, 80% of middle east oil exports are bounded for ASEAN and East Asian countries, making the region largely vulnerable to unforeseen oil disruptions as the transportation route pass-through the strait of Hormoz which is highly vulnerable due to political tensions in the middle east, and another chock point in the Malacca strait in Singapore due to high volume of traffic of vessel pass-through this small strait (Otaka and Han, 2016). At the country context, Cambodia has a very high import dependency on imports of coal, oil (petroleum products), and electricity. Statistics showed that Cambodia’s import dependency increased from 50% to almost 60% during 2013–2016 (MME, 2016). By realising the energy supply risk, Cambodia’s government is making efforts to develop energy infrastructures, such as oil refineries and tapping domestic oil production by 2020. Securing adequate supply via affordable prices and environmentally sustainable uses is the main objective of Cambodia’s overall energy policy. However, Cambodia has experienced rapid growth of energy demand, higher oil import dependence, a growing high share of coal use, and, thus, greater challenges to energy security and managing CO2 emissions (Shigeru and Han, 2019). These could threaten the stable supply of national and local affordable energy. Literally, there are exist relationship between energy efficiency and energy security. Chen et al. (2019) investigated the impact of international sanctions (including unilateral, plurilateral, U.S., EU, UN, economic, and non-economic cases) on energy efficiency, and the found that the imposition of unilateral sanctions leads to a 0.067% decrease in energy efficiency, but that of plurilateral sanctions positively contribute to energy efficiency in the case of the full sample of countries. This also imply energy security and welfare of sanctioned states. Similarly, Mingbo et al. (2018) investigated the long run cointegration relationship between housing price and divorce in China using panel data for 31 provinces over the period 1997–2015. The study found that housing price and divorce have the long run cointegrated relationship in the full samples, in the short run, housing price has a positive effect on divorce in the whole
Energy security has been extensively studied at the national and regional levels, and its meaning changes depending on its scope. From a supply perspective, several indicators are used as a proxy of energy security, such as self-sufficiency ratio, import dependency, diversification
1 EAS is composed of ASEAN plus Australia, China, India, Japan, the Republic of Korea, and New Zealand. United States of America and Russia are invited to join the summit or other important meetings of EAS.
population incurred difficulty affording space heating, and about 80% of the population mentioned that they used less heat than needed in order to make ends meet (Papada and Kaliampakos, 2016). Despite significant progress in providing electricity access to all as stated in the United Nations Sustainable Development Goals (MDG), the world likely falls short of the target set in 2030 to ensure electricity access for all citizens. Currently in 2019, there is about 840 million people do not have access to electricity. It is estimated that there will be about 640 million people still left without access to electricity by 2030, and most of whom live in the developing countries of Asia and sub-Saharan Africa (World Bank, 2019a). In ASEAN countries, access to modern energy remain policy priority as countries like Cambodia, Myanmar, Lao PDR still have low access to electricity in rural areas. While some countries in Southeast Asia such as Malaysia, Singapore, Thailand, and Viet Nam have achieved almost 100% energy access, Cambodia and Myanmar are still struggling to accomplish this (Han, 2015). Data from the World Bank (2019b) show about 50% of Cambodia’s population and 57% of Myanmar’s population had electricity access in 2016. However, the rate of electricity access in Cambodia from 2016 to 2019 has been remarkably achieved as number of households connected to grid electricity reached almost 80% (MME, 2019). However, the challenge remains for rural areas where several parts of remoted areas of Cambodia and Myanmar have almost no electricity at all. This concept of energy security defined in this study strongly links energy security with fundamental human rights as reflected in the 65th UN General Assembly’s resolution declaring 2012 as international year for ‘sustainable energy for all’ (United Nations, 2011). This resolution highlighted the importance of energy services that have a profound effect on productivity, health, education, climate change, food and water security, and communication services. It further said the lack of access to clean, affordable, and reliable energy hinders human, social, and economic development and is a major impediment to achieving the Millennium Development Goals. Politicians’ and decision-makers’ lack of understanding of energy insecurity in terms of inadequate and insufficient household and individual energy consumption could delay energy access to all. Thus, this study defined household energy security as the amount of energy needed to meet the basic needs of daily life of the individual and household in terms of cooking, lighting, washing/cleaning, warming/cooling the house. The fundamental research questions of this study are (i) Who are the households facing energy insecurity in Cambodia? (ii) Does insufficient and inadequate energy consumption of the household affect their welfare? (iii) Does empirical evidence using Cambodia data support the existing literatures? What should be the policy implications from the findings of this research? Thus, this study investigates the above research questions by using the 2015 Cambodia Socio-Economic Survey (CSES), and by defining household energy insecurity and determining how it impacts welfare of the households. The results will help formulate policy implications to strengthen the energy security of the households. The paper is organised as follows. First, the study reviews the current government’s policy toward energy security, and then defines the concept of energy insecurity of the household in Cambodia, the stylised facts of energy insecurity, and then conceptualises the impact of energy insecurity on welfare of the households. Second, the study describes the dataset and variables, followed by the results of the study. Finally, it presents the conclusion and policy implications for strengthening energy security of the households. 2. Literature reviews of energy security and national policies
2
H. Phoumin, F. Kimura
Economic Modelling xxx (xxxx) xxx
Darussalam, stated that the country uses final energy demand as the efficiency indicator and aims at a 10% reduction from the business-as-usual scenario by 2030. The Ministry of Mines and Energy (MME) of Cambodia faced challenges as country’s primary energy consumption grew by double over the past 10 years (Ministry of Mines and Energy of Cambodia (MME), 2019). This strong growth of energy consumption requires the Ministry to take appropriate actions in terms of energy efficiency and conservation. In the recent development in May 2019, the MME with technical support from the Economic Research Institute for ASEAN and East Asia (ERIA) launched the study on Cambodia Energy Efficiency Master Plan (CEEMP). The plan aims to review the existing and needed policies to support the energy efficiency programmes and projects in Cambodia. The plan will use the best practices and will adopt the appropriate energy efficiency road map in commercial & residential sectors, industries and transportation sectors. It is expected CEEMP will guide all programmes and projects towards investments in energy efficiency as the Plan will set the targets of energy savings by sectors which require the serious implementations of energy efficiency in all sectors. The Cambodia Energy Efficiency Master Plan is expected to be completed by early 2020.
country, particularly the Eastern region. Thus, the study suggested that the government should take the short-run housing price regulation to slowdown the divorce rate. However, since the study uses the cross-section data, the household energy efficiency is not available, and its concept of household energy insecurity has defined based on minimum need of energy consumption. 2.2. National policies for energy security at the household level The Power Sector Strategy 1999–2016 focuses on household energy security given its objectives to provide an adequate supply of energy throughout Cambodia at reasonable and affordable prices, and to ensure a reliable and secure electricity supply at appropriate prices for the expansion of Cambodia’s economy. At the present time in 2019, the Prime Minister Hun Sen addressed to the public about a huge shortage of 400 Megawatts (MW) supply of electricity during the dry season from March to June. This kind of unpredictable shortage of electricity supply was due to the weather condition that made low water level of hydropower reservoir (Phnom Penh Post, 2019). This obvious large electricity shortage undermine the level of energy security and it also implies the weak capacity of Electricity of Cambodia (EDC) for managing such a vulnerable power supply mix. The Electricity Law was passed in 2011, with the aim, amongst others, of ensuring the protection of the rights of consumers to receive the reliable and adequate supply of electric power services at reasonable costs. However, Cambodian electricity cost/tariff ($US Cent 18.5/kWh for residential sector whose electricity consumption is over 201 kWh/ month) is the highest in ASEAN (Electricity Authority of Cambodia AEC, 2019). In fact, Cambodian should provide the fare electricity price as most investment cost on power generation such as coal-fired power plants and hydropower are based on conventional technology; while Singapore, the second highest electricity tariff in ASEAN, has internalised the externality cost such as Carbon Dioxide (CO2) into the tariff structure. Other countries such as Thailand, Malaysia and Indonesia has fairly low electricity price, and many new fleets of power generations are based on high technologies such as Ultra-super Critical technology for the coal-fired power plant, or gas-combined cycle for the gas turbines, and yet these countries still can afford to provide relatively low electricity tariff. Cambodian may need to improve the governance system of the electricity sector to ensure that cost can be reduced through the transparent process of cost/tariff structure. Currently, it seems that there is public discomfort about how the Electricity of Cambodia (EDC) manages the supply of electricity (black out very often) together with the high cost of electricity. In 2006, the Royal Government of Cambodia approved the Rural Electrification by Renewable Energy Policy, which acknowledges the Master Plan Study on Rural Electrification by Renewable Energy in the Kingdom of Cambodia as the guiding document for the implementation of its projects and programmes. The Master Plan envisions achieving a 100% level of village electrification, including battery lighting, by 2020, and a 70% level of household electrification with grid-quality electricity by 2030. By 2015, Cambodia had already achieved a 15% level of rural electricity from solar, small hydropower and distributed energy systems. The Master Plan also laid out clear targets, investments, and responsibilities, with 1,828,485 households to be connected to the national grid by 2020. An additional 260,000 households in very remote areas —too far from the planned grid extension—will be supplied via off-grids using fossil fuel and renewable energy (220,000 households) and solar home systems (40,000 households). The total cost for expanding the rural grid is estimated at US$1.37 billion. In the plan, Electricity of Cambodia will be responsible for the overall planning, development, investment, and operation of the rural medium-voltage sub-transmission lines, and it will partner with private rural energy enterprises to expand, operate, and maintain low-voltage distribution and service lines. The energy efficiency and conservation goals submitted to the 5th EAS Energy Ministers Meeting, held on 20 September 2011 in Brunei
3. Empirical framework 3.1. Concept of household energy security vis a vis household energy insecurity From the outset of the study, the concept of the household’s energy security must be defined quantitatively to distinguish the group of households who are energy insecurity vis-a-vis households who are energy security. This study assumes that improving a household’s energy security could also improve income, health, and education. The World Bank (2016) reported that households in developing countries had seen households’ income increases and their children also performed better in school by just having adequate lighting and a reliable supply of electricity in the evening as this electricity necessity can allow children to study homework at night, and facilitate households’ factors of income generation. As mentioned above in the introduction, a household faces energy insecurity if its consumption is insufficient to meet the basic need of daily life in terms of requirements for cooking, lighting, washing/cleaning, warming/cooling the house. In Cambodia, the average annual electricity consumption is about 300 kWh/person (Shigeru and Han, 2019); however, its subsidy policy is applied to any household whose electricity consumption is less than 50 kWh/month or 600 kWh/year. With the average household size of 4.6 in Cambodia, this means that a person can receive electricity subsidy if she/he has annual electricity consumption less than 130 kWh, which is significantly below the average of electricity consumption in Cambodia. United Nations Development Programme (UNDP) (2010) uses a standard of electricity consumption within the range of 50–100 kWh per person per year to sustain life. Similarly, in the case of Sri Lanka, Tennakoon (n.d.) uses the quantitative threshold of per capita annual consumption of 120 kWh electricity for lighting and 35 kg of liquefied propane gas equivalent for cooking. However, some studies use an energy requirement per household cut-off at 2125 kWh by using Guatemala data for 1998–99 (Foster et al., 2000). The quantitative measures defining minimum energy needs are sometimes arbitrary because differences in cooking practices and heating/cooling by region, climate, and culture complicate the level of energy required (Shahidur et al., 2012). Some empirical works defined insufficient energy consumption as a condition where a household’s energy expense is more than 10% of its total income such that it will begin to impact general household welfare. The idea is that when households are forced to spend as much as 10% of income on energy, they are being deprived of other basic goods and services (Barnes, 2010). Based on the UK Department of Energy and Climate Change (DECC, 2015), a household will need to spend more than 10% of their income in order to 3
H. Phoumin, F. Kimura
Economic Modelling xxx (xxxx) xxx
3.3. Dataset, variables and multicollinearity test
acquire sufficient energy to have adequate standard of warmth. This study adopts both approaches of using the 10% threshold of household energy expenditure as an energy insecurity cut-off point, and the minimum 600 kWh/year household electricity consumption threshold. However, using the electricity approach as a cut-off point has a drawback as it is limited to only those houses that are connected to electrical grids.
The study uses the 2015 CSES dataset for analysis. CSES is a survey of households and their members on housing conditions, education, economic activities, production and income, level and structure of consumption, health, victimisation, vulnerability, energy source and consumption, etc. The CSES database at the National Institute of Statistics is open to external researchers for research and analysis; in 2015, it comprised 3840 households. Key variables of interest for the investigation include ‘energy insecurity’ and ‘welfare’, which are the ‘health outcome, school outcome, earning outcome, hours worked of children’ variables. The variable ‘energy insecurity’ is derived from the probability function based on several criteria as described in the methodology. Other household and community characteristics are described in Table 1. The above empirical model employs the multivariate regression models to investigate its hypotheses. The study also performed the multicollinearity test of the independent variables and it found that all the right-hand side variables are acceptable as the Variance Inflation Factor (VIF) is small well below 10 (see Table 2). However, to ensure the fitted coefficient and T-statistics are not inflated, the study used the robustness standard errors correction techniques to correct the estimation of T-statistics.
3.2. Empirical model In this study, we suspected that the energy insecurity of the households is likely to have direct impact on food consumption and on the education expenditure/consumption for children. Based on the aboveconceptualised energy insecurity, we can write the following: PðYi ¼ 1Þ ¼ IðYi ¼ E > 10%Þ if energy insecurity; PðYi ¼ 0Þ if otherwise where E is the share of energy expenditure to the total expenditure, I(.) is an indicator function that takes on a value of {1} if the bracketed expression is true, and a value of {0} if otherwise. Then I(.) equals the value of {1} and the household would be considered energy insecure. Thus, incidence of energy insecurity is:
4. Results and discussions
n 1 X Energy Insecurityi ¼ ½IðYi ¼ E > 10%Þ N i¼1
Based on the provided definition, it is estimated that 27% of the total households remain energy insecure. However, when broken down into on-grid and off-grid, 13% of the households connected to the grid still
A similar approach is applied to household energy insecurity using electricity consumption with the threshold of 50 kWh/month, or 600 kWh/year. Once the energy insecurity indicator has been determined, it is very important to investigate whether it has any impact on households’ welfare. The hypothesis is that household of energy insecurity is believed to substantially impact a households’ welfare including food consumption, education, and health of the individual in the household. To investigate energy insecurity on households’ welfare, two structural equations are constructed. The right-hand side of the equation in the first structural equation uses the independent variable ‘Energy Insecurity’ explicitly as it will affect households’ welfare. The second structural equation, the righthand side variable ‘Share of energy expenditure to total expenditure’, is used to investigate the magnitude of impact. Thus, the model specification can be written as: Welfarei ¼ α0 þ β1 Energy Insecurityi þ β2 Xi þ β3 Rurali þ Ui1
Table 1 Description of variables. Description of variables
Mean statistics
Log household food consumption
Logarithm of household food consumption in the past 12 months (KR) Logarithm of household education/ expenditure in the past 12 months (KR) Share of energy expenditure to total expenditure (Percentage)
15.99,678
Log household education consumption/ expenditure Share of energy expenditure to total expenditure Household size Log household’s income
(1)
Electricity access Without piped water access
Welfarei ¼ α0 þ β1 Share of Energy Expenditurei þ β2 Xi þ β3 Rurali þ Ui1 (2)
Without tubed well access
where ‘welfare’ is the dependent variable representing a household’s food consumption, education consumption/expenditure, child’s schooling outcome, and health outcomes. The proxy variable ‘schooling outcome’ is an index of Schooling Attainment Relative to Age (SAGE) variable (Han, 2015). The SAGE index is simply derived from the equation below:
Household head with no education Household head with primary education incomplete Household head with primary education complete Household head with upper secondary education
SAGE ¼
Name of variable
Years of schooling 100 Age Age of School Entry
Household head with university Electricity consumption per capita Female household head
The higher the index, the better the schooling outcome of children. The variable Xi is the set of exogenous variables representing the household’s characteristics such as household’s income, electricity access, electricity consumption per capita, education, and access to clean water. The variable Rurali is the community characteristic if the household is residing in the rural or otherwise.
Rural
6.129,855
Household size Logarithm of household’s income in the past 12 months ¼ 1 if household has electricity access; ¼ 0 for otherwise ¼ 1 if without piped water access; ¼ 0 for otherwise ¼ 1 if without tubed well access; ¼ 0 for otherwise ¼ 1 if household head with no education; ¼ 0 for otherwise ¼ 1 if household head with primary education incomplete; ¼ 0 for otherwise ¼ 1 if household head with primary education complete; ¼ 0 for otherwise
4.506,642 16.4995
¼ 1 if household head with upper secondary education; ¼ 0 for otherwise ¼ 1 if household head with university degree; ¼ 0 for otherwise Electricity consumption per capita (kWh/capita/year) ¼ 1 if female household head; ¼ 0 for otherwise ¼ 1 if rural; ¼ 0 for otherwise
.0255,275
KR ¼ Cambodian riel. Source: 2015 Cambodia Socio-Economic Survey. 4
13.39,273
.7272017 .0080,771 .2342366 .1802553 .4464704
.1875488
.041,417 252.5145 .756,968 .5965095
H. Phoumin, F. Kimura
Economic Modelling xxx (xxxx) xxx
Table 2 Multicollinearity test.
Table 4 Regression coefficient estimates of food consumption.
Variable
VIF
1/VIF
Household head with primary education complete Household head with no education Household head with primary education incomplete Electricity consumption per capita Rural Log household’s income Energy insecurity of household Household size Household head with university Electricity access Household head with upper secondary education Female household head Without tubed well access Without piped water access Mean VIF
2.82 2.39 2.12 1.59 1.53 1.44 1.34 1.34 1.34 1.32 1.19 1.13 1.12 1.01 1.55
0.354,838 0.417,612 0.471,108 0.629,142 0.655,284 0.692,652 0.746,116 0.746,855 0.748,552 0.756,013 0.840,303 0.888,105 0.888,915 0.991,721
Ongrid Offgrid Total
Total
914 87.30% 1898 68.00% 2812 73.27%
133 12.70% 893 32.00% 1026 26.73%
1047 100% 2791 100% 3838 100%
Coefficient
Coefficient
Energy insecurity of household Share of energy expenditure to total expenditure Household size
None
-.3871497*** (.0321,373) None
Without piped water access Without tubed well access Household head with no education Household head with primary education incomplete Household head with primary education complete Household head with upper secondary education Household head with university Electricity consumption per capita Female household head Rural Constant
-.0413,496*** (.0081,013)
Multiple regression model using Robust Standard Error Correctionb
.126,217*** (.0050,765) .0720,751*** (.0093,409) .0819,193*** (.0145,262) -.1786928*** (.0529,703) -.0935,054*** (.0133,904) -.1081735*** (.0220,311) -.0766,681*** (.0196,922)
.1115705*** (.0047,435) .1010312*** (.0095,474) .0762,213*** (.0150,818) -.2073098** (.0637,238) -.1066915*** (.0140,031) -.1493641*** (.023,567) -.1048406*** (.0207,686)
-.0120,718 (.0202,931)
-.0290,963 (.0223,665)
-.0264,382 (.0387,978)
-.0040,873 (.0420,527)
.0626,308** (.036,567) .0009257*** (.000142) .0864,377*** (.0139,653) -.137,467*** (.0144,018) 14.29,294*** (.1479394)
.1151599** (.0353,296) .0003998*** (.0000859) .0854,652*** (.0144,705) -.1628727*** (.0160,552) 13.84,663*** (.1443285)
Goodness of fit. Number of obs. ¼ 1155; F(14, 1140) ¼ 26.57; Prob > F ¼ 0.0000; Rsquared ¼ 0.2471; Root MSE ¼ 1.105. Note: (*) represents the significant level at 10%; (**) represents the significant level at 5%; and (***) represents the significant level at 1%. Number in bracket is the robust standard error. a The regression model includes the independent variable ‘share of energy expenditure to total expenditure’. b The regression model includes the independent variable ‘energy insecurity of the household’. Source: Author’s calculation.
The impact of a household’s income on welfare such as food consumption and education expenditure for children: The coefficient estimates of a household’s income show that food expenditure is likely to increase by about 7% for every doubling of a household’s income; similarly, education consumption and expenditure is expected to increase by about 17%. Thus, the impact of a household’s income has a direct benefit to welfare of the household. The impact of household head’s education on welfare such as food consumption and education’s expenditure for children: The coefficient estimates of household head’s completion of university has a positive impact on both food consumption (a 6% increase), and a positive impact on education expenditure/consumption (a 20% increase). The impact of other household and community characteristics on welfare such as food consumption and education’s expenditure for children: Among other household characteristics, its size has direct impact on food consumption and education of children. The characteristics of the community such as households living in rural area seems to have negative impact on food consumption as well as education
Table 3 Prevalence of household energy insecurity by off-grid and on-grid electricity. Households of energy insecurity (frequency and percentage)
Multiple regression model using Robust Standard Error Correctiona
Log household’s income Electricity access
face energy insecurity, while 32% of households living off-grid face energy insecurity (Table 3). To investigate the main hypothesis, the study regresses the households’ food consumption and education expenditure/consumption on the share of energy expenditure to the total expenditure and also on sets of household and community characteristics (Tables 4, 5 and 6). The result of coefficient estimates can be interpreted as follows: The impact of energy insecurity on welfare such as food consumption and education expenditure for the children: The coefficient estimates of share of energy expenditure to total expenditure’ in Table 3 showed that a household will likely reduce its food consumption by about 4.1% for every 1% increase in the share of energy expenditure to total expenditure. This means that every percentage point of energy expenditure will impact a household’s food consumption substantially. If the data allowed, this will further impact general household health, especially the children and elderly persons living in the household. Likewise, the coefficient estimates of ‘share of energy expenditure to total expenditure’ in Table 5 showed that households will likely reduce their education consumption/expenditure by about 5.8% for every 1% increase in the share of energy expenditure to total expenditure/consumption. This means that every percentage point of energy expenditure will impact the household’s education consumption/expenditure substantially. Again, energy insecurity likely reduces a household’s food consumption by about 38% compared to households that are energy secure. Likewise, energy insecure households are also likely to have reduced their education expenditure by 26% as compared to energy secure households. Thus, the impacts of energy insecurity are enormous as they affect direct food consumption and education expenditure, which are fundamental to support growth and human capital formation of children. It is believed that the less spending on education and less food consumption in households will further impact on children’s performance at school, especially their human capital formation. Therefore, the regression results in Table 6 shows that every percentage increase of ‘share of energy expenditure to the total expenditure’ is likely to reduce the children’s SAGE index by 47%.
Households of energy security (frequency and percentage)
Dependent variable ‘Log household food consumption’
Source: Author’s calculation. 5
H. Phoumin, F. Kimura
Economic Modelling xxx (xxxx) xxx
Table 5 Regression coefficient estimates of education consumption/expenditure. Dependent variable ‘Log household education consumption/ expenditure’
Energy insecurity of household Share of energy expenditure to total expenditure Household size Log household’s income Electricity access Without piped water access Without tubed well access Household head with no education Household head with primary education incomplete Household head with primary education complete Household head with upper secondary education Household head with university Electricity consumption per capita Female household head Rural Constant
Multiple regression model using Robust Standard Error Correctiona
Multiple regression model using Robust Standard Error Correctionb
Coefficient
Coefficient
None
-.2621564** (.124,238) None
-.0585,014*** (.0130,137)
Table 6 Regression coefficient estimates of schooling attainment relative to age (SAGE).
.0965,153*** (.0278,622) .1666177*** (.0484,125) .2905582*** (.1107727) .2213673 (.3583839) .0945,942 (.0918,691) -.6045892*** (.1477813) -.3900032*** (.0949,524)
.079,152** (.0266,817) .2035521*** (.0455,854) .2782467** (.1121011) .1756231 (.3712069) .0902,604 (.0931,144) -.6577931*** (.1492502) -.4435312*** (.0968,051)
-.0935,535 (.0977,389)
-.1159162 (.100,138)
-.0713,393 (.1437776)
-.0661,887 (.1488217)
.2170157* (.1482588) .0015,156*** (.0004294) .1336658 (.0829,068) -.5287331*** (.078,998) 10.29,859*** (.7414394)
.3188259** (.1445511) .0007297** (.0002679) -.1501823* (.0850,642) -.5313*** (.0800,466) 9.711,694*** (.712,983)
Dependent variable ‘SAGE’ - An index of schooling attainment relative to age
Multiple regression model using Robust Standard Error Correctiona
Multiple regression model using Robust Standard Error Correctionb
Coefficients
Coefficients
Energy insecurity
None
Share of energy expenditure to total expenditure Without piped water access Without tubed well access
-.4792831*** (.1270687)
2.06767** (1.049,695) None
Household head with no education Household head with primary education incomplete Household head with primary education complete Household head with upper secondary education Household head with university Electricity consumption per capita Female household head Rural Constant
1.160,382 (5.151,614) 2.150,683 (1.365,153) 13.5516*** (1.971,471) 9.285,921*** (1.562,142)
1.407,132 (5.177,143) 1.960,529 (1.36,468) 14.24,124*** (1.973,232) 9.720,728*** (1.563,496)
1.226,556 (1.690,172)
1.444,921 (1.70,026)
2.440,661 (2.553,496)
2.665,434) (2.58,741
11.25,608*** (2.553,982) .014,578*** (.0030,274) .3292066 (1.288,788) 4.979,571*** (1.212,425) 82.81,464*** (1.957,626)
12.80,953*** (2.523,761) .0067,002** (.0029,038) .3081961 (1.288,737) 5.220,719*** (1.227,291) 82.4632*** (1.983,757)
Goodness of fit. Number of obs. ¼ 3837; F(14, 3822) ¼ 208.74; Prob > F ¼ 0.0000; Rsquared ¼ 0.4795; Root MSE ¼ 0.3566. Note: (*) represents the significant level at 10%; (**) represents the significant level at 5%; and (***) represents the significant level at 1%. Number in bracket is the robust standard error. a The regression model includes the independent variable ‘share of energy expenditure to total expenditure’. b The regression model includes the independent variable ‘energy insecurity of the household’. Source: Author’s calculation.
Goodness of fit. Number of obs. ¼ 1155; F(14, 1140) ¼ 27.92; Prob > F ¼ 0.0000; Rsquared ¼ 0.2739; Root MSE ¼ 1.0851. Note: (*) represents the significant level at 10%; (**) represents the significant level at 5%; and (***) represents the significant level at 1%. Number in bracket is the robust standard error. a The regression model includes the independent variable ‘share of energy expenditure to total expenditure’. b The regression model includes the independent variable ‘energy insecurity of the household’. Source: Author’s calculation.
5. Conclusion and policy implications Energy security has been widely known from the supply side perspective where leaders try to secure energy supplies and also reduce the dependency on fossil fuel imports. Generally, leaders try to build an energy policy framework that shields countries from supply disruption risks. However, energy insecurity occurs when households are deprived of the energy needed to meet the basic minimum requirements for cooking, lighting, heating/cooling. Since the concept of energy insecurity fluctuates, the study adopted the share of energy consumption being 10% to the total expenditure as the threshold of household energy insecurity. With this definition, the national incidence of energy insecurity is estimated at about 27% of the total households in Cambodia. The study further separates the incidence of energy insecurity between off-grid and on-grid households. The result shows about 13% of the households connected to the grid still face energy insecurity; about 32% of households living off-grid face energy insecurity. The major findings about household food consumption show that it will likely decrease by about 4.1% for every 1% increase in the share of energy expenditure to total expenditure, with education consumption/ expenditure similarly being reduced by about 5.8%. Energy insecure households consume about 38% less than energy secure ones and will
expenditure. This result indicates that households spent less on food and education of children if they lived in rural or remote areas. The result also points to the fact that households living in rural areas are engaged in a subsistence economy where their labour is not counted and are making their own living by farming. The results also show that a household head having finished a university degree has substantial positive impacts on a household’s food consumption and education expenditure. Likewise, a household head with just a primary education or less has negative impacts on food consumption and education expenditure. This is self-evident because education is crucial to household achieving household income, and it further affects purchasing power. Finally, other community characteristics such as electricity access, and access to water sources such as piped or tubed well are keys for a positive impact on welfare of the households. The access to electricity and water sources represent the impact of infrastructure on welfare.
6
H. Phoumin, F. Kimura
Economic Modelling xxx (xxxx) xxx
likely spend 26% less on education. Thus, the impacts of energy insecurity are enormous. The study further found that every percentage point increase in the share of energy expenditure to the total expenditure is likely to reduce children’s school performance by 47%. For every doubling of a household’s income, the study found that it increases food expenditure by about 7%, with education consumption and expenditure similarly expected to increase by about 17%. Thus, the impact of household’s income has direct benefit to welfare of the household. The household-head’s completion of university has positive impact on both food consumption (a 6% increase), and education expenditure/consumption of children (a 20% increase). Among other household characteristics, its size has direct impact on food consumption and education expenditure, while residency in rural areas seems to have negative impact. The other community characteristics such as access to electricity and water sources represent the positive impact of infrastructure on household welfare. Thus, this study confirmed that energy insecurity has enormous negative impact on households and the human capital formation of the children. The above findings imply the following:
ASEAN Centre for Energy [ACE], 2019. Renewable energy: outcome-based Strategy. http://www.aseanenergy.org/programme-area/re/. (Accessed 8 September 2019). Barnes, D., 2010. The concept of energy poverty,’ energy for development and poverty reduction, 19 June. http://www.energyfordevelopment.com/2010/06/energy-pove rty.html. (Accessed 27 February 2019). Chen, Yin E., Fu, Qiang, Chang, Chun-Ping, 2019. International Sanctions’ Impact on Energy Efficiency in Target States, Economic Modelling. Economic Modelling, 23 July 2019. Dept for Energy and Climate Change (DECC) (2015), 2015. Annual Fuel Poverty Statistics Report. DECC, London. Electricity Authority of Cambodia [AEC], 2019. Electricity tariff- in Khmer language. https://www.eac.gov.kh/document/tariffdecidelist. (Accessed 10 September 2019). Foster, V., Tre, J.-P., Wodon, Q., 2000. Energy Prices, Energy Efficiency, and Fuel Poverty. World Bank, Washington, DC. Han, P., 2015. Energy Security and Sustained Growth: Analysis of Energy Outlook and Savings Potential in the EAS Region. Asian Development Bank Institute, Tokyo. https ://www.asiapathways-adbi.org/2015/11/energy-security-and-sustained-grow th-analysis-of-the-energy-outlook-and-savings-potential-in-the-eas-region/. (Accessed 18 February 2019). Han, P., Kimura, F., 2019. Cambodia’s energy poverty and its effects on social wellbeing: empirical evidence and policy implications. Energy Policy 132, 283–289. International Energy Agency (IEA), 2017. Southeast Asia Energy Outlook 2017: from Poverty to Prosperity. World Energy Outlook Special Report. IEA, Paris. Kutani, I., 2013. Study on the Development of an Energy Security Index and an Assessment of Energy Security Policy for East Asia Countries. Economic Research Institute for ASEAN and East Asia, Jakarta. http://www.eria.org/publications /study-on-the-development-of-an-energy-security-index-and-an-assessment-of-en ergy-security-for-east-asian-countries/. (Accessed 18 February 2019). Mingbo, Zheng, Chen, Yin E., Feng, Gen-Fu, Wen, Jun, Chang, Chun-Ping, 2018. Divorce and housing price in 31 provinces of China. Bulletin of Monetary Economics & Banking 21 (3), 173–188. Ministry of Mines and Energy of Cambodia (MME), 2016. Cambodia National Energy Statistics 2016. Economic Research Institute for ASEAN and East Asia, Jakarta. http: //www.eria.org/publications/cambodia-national-energy-statistics-2016/. (Accessed 19 February 2019). Ministry of Mines and Energy of Cambodia (MME), 2019. Cambodia Basic Energy Plan. Economic Research Institute for ASEAN and East Asia, Jakarta. http://www.eria. org/publications/cambodia-basic-energy-plan/. (Accessed 26 March 2019). Otaka, Y., Han, P., 2016. Study on the Strategic Usage of Coal in the EAS Region: A Technical Potential Map and Update of the First-Year Study. Economic Research Institute for ASEAN and East Asia, Jakarta. http://www.eria.org/publications/s tudy-on-the-strategic-usage-of-coal-in-the-eas-region-a-technical-potential-map-andupdate-of-the-first-year-study/. (Accessed 18 February 2019). Papada, L., Kaliampakos, D., 2016. Measuring energy poverty in Greece. Energy Policy 94, 157–165. Phnom Penh Post, 2019. Kingdom Lacks up to 400 MW in Available Electricity. https://www.phnompenhpost.com/national/kingdom-lacks-400mw-available-electri city. (Accessed 9 September 2019). Shahidur, K., Barnes, D., Samad, H., 2012. Are the energy poor also income poor? Evidence from India. Energy Policy 47, 1–12. Shigeru, K., Han, P., 2019. Energy Outlook and Energy Saving Potential in East Asia 2019. Economic Research Institute for ASEAN and East Asia, Jakarta. http://www.eria.org /publications/energy-outlook-and-energy-saving-potential-in-east-asia-2019/. (Accessed 26 March 2019). Cambodia Socio-Economic Survey (CSES), 2015. Cambodia Socio-Economic Survey. National Institute of Statistics, Cambodia. Ministry of Planning. The ASEAN Post, 2018. Southeast Asia’s Hope for Renewable Energy in the Year Ahead. https://theaseanpost.com/article/southeast-asias-hope-renewable-energy-year-ahea d. (Accessed 9 September 2019). United Nations, 2011. 2012 International Year for Sustainable Energy for All. htt p://www.un.org/en/events/sustainableenergyforall/. (Accessed 19 February 2019). United Nations Development Programme (UNDP), 2010. Energy for a Sustainable Future: the Secretary-General’s Advisory Group on Energy and Climate Change Summary Report and Recommendation. UNDP, New York. US Energy Information Administration (EIA), 2018. Residential energy consumption survey. https://www.eia.gov/consumption/residential/. (Accessed 22 February 2019). World Bank, 2016. Sustainable Development Goal on Energy (SDG7) and the World Bank Group. http://www.worldbank.org/en/topic/energy/brief/sustainable-develop ment-goal-on-energy-sdg7-and-the-world-bank-group. (Accessed 21 February 2019). World Bank, 2019a. More People Have Access to Electricity than Ever before, but World Falling Short of Sustainable Energy Goals. https://www.worldbank.org/en/news /press-release/2019/05/22/tracking-sdg7-the-energy-progress-report-2019. (Accessed 10 September 2019). World Bank, 2019b. Access to Electricity: % of Population. https://data.worldbank .org/indicator/EG.ELC.ACCS.ZS. (Accessed 21 February 2019).
o Welfare impacts: policymakers will need to identify households whose energy consumption is insufficient to meet basic needs. Once they are identified, the appropriate energy policy should target those vulnerable households to support them to have sufficient energy. Electricity access is the first priority, followed by a subsidy. In addition, the policy needs to address off-grid and on-grid families differently. o Infrastructure: Access to electricity, piped water systems, and other infrastructure are fundamental to increase households’ opportunities and income. Nonetheless, access to electricity is the first step for households to consume clean electricity in comparison to biomass or kerosene for heating and lighting. Thus, policymakers will need to find appropriate policies to ensure that households are connected to the electrical grid as fast as possible. While the development of the power grid and distribution will remain slow due to investment patterns, other measures, such as rooftop solar photovoltaics, solar farms, and stand-alone small generators will provide fast access to remote areas. Policies to support these distributed energy systems will need to be designed to support the investment in this area. o The government may need to review the current policy of subsidies for households whose electricity consumption is below 50 kWh/ month, as this threshold is too low to meet basic needs. The threshold could be increasing up to double in the first step and check how it impacts the welfare of the household. Thus, further study will be needed to see how much subsidy the government can afford, and to what extent the subsidy could have positive long-term impact on household welfare. Acknowledgements This study is carried out under academic project funding from Economic Research Institute for ASEAN and East Asia [ERIA]. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.econmod.2019.09.024. References ASEAN, 2007. Cebu declaration on East Asian energy security. Cebu, Philippines, 15 January. https://asean.org/?static_post¼cebu-declaration-on-east-asian-energy-secu rity-cebu-philippines-15-january-2007-2. (Accessed 18 February 2019).
7
Contents lists available at ScienceDirect
Economic Modelling journal homepage: www.journals.elsevier.com/economic-modelling
The impacts of energy insecurity on household welfare in Cambodia: Empirical evidence and policy implications Han Phoumin a, *, Fukunari Kimura b a b
Energy Economist with Economic Research Institute for ASEAN and East Asia (ERIA) Based in Jakarta, Indonesia Keio University, Chief Economist with Economic Research Institute for ASEAN and East Asia (ERIA) Based in Jakarta, Indonesia
A R T I C L E I N F O
A B S T R A C T
JEL classification: Q48 C39 I39
This study investigates the impacts of energy insecurity on household welfare in Cambodia. The notion of energy insecurity is not well understood in the literature as well as in local contexts. This study defines household energy insecurity as the status quo derived from the interplay of inadequate and insufficient energy consumption that prevents households from meeting basic energy needs. The notion of energy insecurity can only be well understood by investigation in the local context as it varies from place to place. Households facing insufficient energy consumption may forgo many other opportunities. Once energy security has been defined in the Cambodian context, the study employs multiple regression models using the Cambodia Socio-Economic Survey Data (2015) to investigate the impacts of household energy insecurity. The study confirmed that energy insecurity has enormous negative impact on welfare of the households, with a further negative impact on the human capital formation of the children. The findings will lead to policy implications to improve household energy security, and thus impact economic, social, and environmental development.
Keywords: Energy insecurity Schooling and welfare
1. Introduction The concept of household’s energy security is the 1st step to be defined to get general understanding of its meaning and scope. In the literatures, the word “security” means “the state of being free from danger or threat”. Adopting this definition, the household’s energy security takes to mean the interplay of adequate and sufficient consumption of households to meet basic energy needs. In other word, the household is energy insecurity when that household is in the state of inadequate and insufficient energy consumption to meet basic energy needs. As energy uses, especially the electricity and liquefied Natural Gas (LPG) is centrally important for supporting the daily life of households in terms of cooking, lighting, reading, watching television, and other household income generating business activities, the inadequate and insufficient consumption of these energy will affect their quality of life at various degrees. Let alone, about 65 million people of Association of Southeast Asia Nations (ASEAN) has denied access to electricity and about 250 million people relies on biomass as a cooking fuel (IEA, 2017). Some countries in ASEAN such as Cambodia, Lao PDR and Myanmar, remote islands of Indonesia and Philippines, are facing challenges in providing electricity access in affordable price. Households having access to grid electricity is not the end of the problem as they face high
electricity tariff. For the better-off households the high electricity tariff may not impact much for their affordability of electricity consumption, but for the poor households, the electricity tariff may impact their affordability, leading to insufficient electricity consumption. Han and Kimura (2019) confirmed that access to electricity alone does not tell about adequate energy consumption. People with less wealth and income tend to use electricity only for important activities such as during dining and lighting for a few hours at night, leading to insufficient consumption. Energy insecurity exists not only in developing countries, but in developed countries as well; for instance, a study conducted by the US Energy Information Administration (EIA, 2018) found that about one-third of US households struggled to pay energy bills just to maintain adequate heating and cooling in their homes in 2015. According to the same report, about one in five households reported reducing or forgoing necessities such as food and medicine to pay an energy bill, and 14% reported receiving a disconnection notice for energy service. Households may also use less energy than they would prefer; 11% of households surveyed reported keeping their home at an unhealthy or unsafe temperature. In Europe, 50 million–125 million people in 2009 were struggling to meet their energy needs, including in some countries having undergone economic austerity reforms; for example, since the outbreak of the economic crisis of Greece in 2009, 64% of the country’s northern
* Corresponding author. E-mail addresses: [email protected] (H. Phoumin), [email protected] (F. Kimura). https://doi.org/10.1016/j.econmod.2019.09.024 Received 28 May 2019; Received in revised form 16 September 2019; Accepted 17 September 2019 Available online xxxx 0264-9993/© 2019 Published by Elsevier B.V.
Please cite this article as: Phoumin, H., Kimura, F., The impacts of energy insecurity on household welfare in Cambodia: Empirical evidence and policy implications, Economic Modelling, https://doi.org/10.1016/j.econmod.2019.09.024
H. Phoumin, F. Kimura
Economic Modelling xxx (xxxx) xxx
2.1. Supply security of energy
of import source countries, and strategic oil reserves. Most research studies on energy security refer to ‘securing of the amount of energy required for people’s lives, economic, social, and defense activities, among other purposes, at an affordable prices’ (Kutani, ed., 2013). In the East Asia Summit (EAS) context,1 the leaders adopted the Cebu Declaration in 2007 reaffirming the commitment to energy security by working closely together to reduce dependence on conventional fuels through intensified energy efficiency and conservation programmes and expansion of renewable energy systems and biofuel production/utilization (ASEAN, 2007). In recent years, ASEAN has striking towards diversifying its energy mix, and agreed on aspirational target to increase the share of renewable energy to 23% by 2025 in the primary energy mix (ACE, 2019). In recent development in the past few years, countries have been accelerating the installation of solar farm as well as the introduction of solar roof-top in many countries of ASEAN. Thanks for the speed of cost reduction of solar module cost and the gradual shift of know-how capacity to handle the solar and wind technologies into the ASEAN (The ASEAN Post, 2018). However, despite huge attempt and efforts by the leaders in ASEAN to ensure electricity access to all population, there remain challenges in some countries of ASEAN where electricity access remain low and electricity cost is high, and thus making both accessibility and affordability become core economic disadvantage as well as threatening the energy security of the household as these accessibility and affordability threatening daily life of people (Han and Kimura, 2019). From a broad view, the predicted primary and final energy demand from 2015 to 2040, which is expected to grow almost double in the EAS region during the forecasting periods, is posing an increasing threat to energy security (Han, 2015). Further, 80% of middle east oil exports are bounded for ASEAN and East Asian countries, making the region largely vulnerable to unforeseen oil disruptions as the transportation route pass-through the strait of Hormoz which is highly vulnerable due to political tensions in the middle east, and another chock point in the Malacca strait in Singapore due to high volume of traffic of vessel pass-through this small strait (Otaka and Han, 2016). At the country context, Cambodia has a very high import dependency on imports of coal, oil (petroleum products), and electricity. Statistics showed that Cambodia’s import dependency increased from 50% to almost 60% during 2013–2016 (MME, 2016). By realising the energy supply risk, Cambodia’s government is making efforts to develop energy infrastructures, such as oil refineries and tapping domestic oil production by 2020. Securing adequate supply via affordable prices and environmentally sustainable uses is the main objective of Cambodia’s overall energy policy. However, Cambodia has experienced rapid growth of energy demand, higher oil import dependence, a growing high share of coal use, and, thus, greater challenges to energy security and managing CO2 emissions (Shigeru and Han, 2019). These could threaten the stable supply of national and local affordable energy. Literally, there are exist relationship between energy efficiency and energy security. Chen et al. (2019) investigated the impact of international sanctions (including unilateral, plurilateral, U.S., EU, UN, economic, and non-economic cases) on energy efficiency, and the found that the imposition of unilateral sanctions leads to a 0.067% decrease in energy efficiency, but that of plurilateral sanctions positively contribute to energy efficiency in the case of the full sample of countries. This also imply energy security and welfare of sanctioned states. Similarly, Mingbo et al. (2018) investigated the long run cointegration relationship between housing price and divorce in China using panel data for 31 provinces over the period 1997–2015. The study found that housing price and divorce have the long run cointegrated relationship in the full samples, in the short run, housing price has a positive effect on divorce in the whole
Energy security has been extensively studied at the national and regional levels, and its meaning changes depending on its scope. From a supply perspective, several indicators are used as a proxy of energy security, such as self-sufficiency ratio, import dependency, diversification
1 EAS is composed of ASEAN plus Australia, China, India, Japan, the Republic of Korea, and New Zealand. United States of America and Russia are invited to join the summit or other important meetings of EAS.
population incurred difficulty affording space heating, and about 80% of the population mentioned that they used less heat than needed in order to make ends meet (Papada and Kaliampakos, 2016). Despite significant progress in providing electricity access to all as stated in the United Nations Sustainable Development Goals (MDG), the world likely falls short of the target set in 2030 to ensure electricity access for all citizens. Currently in 2019, there is about 840 million people do not have access to electricity. It is estimated that there will be about 640 million people still left without access to electricity by 2030, and most of whom live in the developing countries of Asia and sub-Saharan Africa (World Bank, 2019a). In ASEAN countries, access to modern energy remain policy priority as countries like Cambodia, Myanmar, Lao PDR still have low access to electricity in rural areas. While some countries in Southeast Asia such as Malaysia, Singapore, Thailand, and Viet Nam have achieved almost 100% energy access, Cambodia and Myanmar are still struggling to accomplish this (Han, 2015). Data from the World Bank (2019b) show about 50% of Cambodia’s population and 57% of Myanmar’s population had electricity access in 2016. However, the rate of electricity access in Cambodia from 2016 to 2019 has been remarkably achieved as number of households connected to grid electricity reached almost 80% (MME, 2019). However, the challenge remains for rural areas where several parts of remoted areas of Cambodia and Myanmar have almost no electricity at all. This concept of energy security defined in this study strongly links energy security with fundamental human rights as reflected in the 65th UN General Assembly’s resolution declaring 2012 as international year for ‘sustainable energy for all’ (United Nations, 2011). This resolution highlighted the importance of energy services that have a profound effect on productivity, health, education, climate change, food and water security, and communication services. It further said the lack of access to clean, affordable, and reliable energy hinders human, social, and economic development and is a major impediment to achieving the Millennium Development Goals. Politicians’ and decision-makers’ lack of understanding of energy insecurity in terms of inadequate and insufficient household and individual energy consumption could delay energy access to all. Thus, this study defined household energy security as the amount of energy needed to meet the basic needs of daily life of the individual and household in terms of cooking, lighting, washing/cleaning, warming/cooling the house. The fundamental research questions of this study are (i) Who are the households facing energy insecurity in Cambodia? (ii) Does insufficient and inadequate energy consumption of the household affect their welfare? (iii) Does empirical evidence using Cambodia data support the existing literatures? What should be the policy implications from the findings of this research? Thus, this study investigates the above research questions by using the 2015 Cambodia Socio-Economic Survey (CSES), and by defining household energy insecurity and determining how it impacts welfare of the households. The results will help formulate policy implications to strengthen the energy security of the households. The paper is organised as follows. First, the study reviews the current government’s policy toward energy security, and then defines the concept of energy insecurity of the household in Cambodia, the stylised facts of energy insecurity, and then conceptualises the impact of energy insecurity on welfare of the households. Second, the study describes the dataset and variables, followed by the results of the study. Finally, it presents the conclusion and policy implications for strengthening energy security of the households. 2. Literature reviews of energy security and national policies
2
H. Phoumin, F. Kimura
Economic Modelling xxx (xxxx) xxx
Darussalam, stated that the country uses final energy demand as the efficiency indicator and aims at a 10% reduction from the business-as-usual scenario by 2030. The Ministry of Mines and Energy (MME) of Cambodia faced challenges as country’s primary energy consumption grew by double over the past 10 years (Ministry of Mines and Energy of Cambodia (MME), 2019). This strong growth of energy consumption requires the Ministry to take appropriate actions in terms of energy efficiency and conservation. In the recent development in May 2019, the MME with technical support from the Economic Research Institute for ASEAN and East Asia (ERIA) launched the study on Cambodia Energy Efficiency Master Plan (CEEMP). The plan aims to review the existing and needed policies to support the energy efficiency programmes and projects in Cambodia. The plan will use the best practices and will adopt the appropriate energy efficiency road map in commercial & residential sectors, industries and transportation sectors. It is expected CEEMP will guide all programmes and projects towards investments in energy efficiency as the Plan will set the targets of energy savings by sectors which require the serious implementations of energy efficiency in all sectors. The Cambodia Energy Efficiency Master Plan is expected to be completed by early 2020.
country, particularly the Eastern region. Thus, the study suggested that the government should take the short-run housing price regulation to slowdown the divorce rate. However, since the study uses the cross-section data, the household energy efficiency is not available, and its concept of household energy insecurity has defined based on minimum need of energy consumption. 2.2. National policies for energy security at the household level The Power Sector Strategy 1999–2016 focuses on household energy security given its objectives to provide an adequate supply of energy throughout Cambodia at reasonable and affordable prices, and to ensure a reliable and secure electricity supply at appropriate prices for the expansion of Cambodia’s economy. At the present time in 2019, the Prime Minister Hun Sen addressed to the public about a huge shortage of 400 Megawatts (MW) supply of electricity during the dry season from March to June. This kind of unpredictable shortage of electricity supply was due to the weather condition that made low water level of hydropower reservoir (Phnom Penh Post, 2019). This obvious large electricity shortage undermine the level of energy security and it also implies the weak capacity of Electricity of Cambodia (EDC) for managing such a vulnerable power supply mix. The Electricity Law was passed in 2011, with the aim, amongst others, of ensuring the protection of the rights of consumers to receive the reliable and adequate supply of electric power services at reasonable costs. However, Cambodian electricity cost/tariff ($US Cent 18.5/kWh for residential sector whose electricity consumption is over 201 kWh/ month) is the highest in ASEAN (Electricity Authority of Cambodia AEC, 2019). In fact, Cambodian should provide the fare electricity price as most investment cost on power generation such as coal-fired power plants and hydropower are based on conventional technology; while Singapore, the second highest electricity tariff in ASEAN, has internalised the externality cost such as Carbon Dioxide (CO2) into the tariff structure. Other countries such as Thailand, Malaysia and Indonesia has fairly low electricity price, and many new fleets of power generations are based on high technologies such as Ultra-super Critical technology for the coal-fired power plant, or gas-combined cycle for the gas turbines, and yet these countries still can afford to provide relatively low electricity tariff. Cambodian may need to improve the governance system of the electricity sector to ensure that cost can be reduced through the transparent process of cost/tariff structure. Currently, it seems that there is public discomfort about how the Electricity of Cambodia (EDC) manages the supply of electricity (black out very often) together with the high cost of electricity. In 2006, the Royal Government of Cambodia approved the Rural Electrification by Renewable Energy Policy, which acknowledges the Master Plan Study on Rural Electrification by Renewable Energy in the Kingdom of Cambodia as the guiding document for the implementation of its projects and programmes. The Master Plan envisions achieving a 100% level of village electrification, including battery lighting, by 2020, and a 70% level of household electrification with grid-quality electricity by 2030. By 2015, Cambodia had already achieved a 15% level of rural electricity from solar, small hydropower and distributed energy systems. The Master Plan also laid out clear targets, investments, and responsibilities, with 1,828,485 households to be connected to the national grid by 2020. An additional 260,000 households in very remote areas —too far from the planned grid extension—will be supplied via off-grids using fossil fuel and renewable energy (220,000 households) and solar home systems (40,000 households). The total cost for expanding the rural grid is estimated at US$1.37 billion. In the plan, Electricity of Cambodia will be responsible for the overall planning, development, investment, and operation of the rural medium-voltage sub-transmission lines, and it will partner with private rural energy enterprises to expand, operate, and maintain low-voltage distribution and service lines. The energy efficiency and conservation goals submitted to the 5th EAS Energy Ministers Meeting, held on 20 September 2011 in Brunei
3. Empirical framework 3.1. Concept of household energy security vis a vis household energy insecurity From the outset of the study, the concept of the household’s energy security must be defined quantitatively to distinguish the group of households who are energy insecurity vis-a-vis households who are energy security. This study assumes that improving a household’s energy security could also improve income, health, and education. The World Bank (2016) reported that households in developing countries had seen households’ income increases and their children also performed better in school by just having adequate lighting and a reliable supply of electricity in the evening as this electricity necessity can allow children to study homework at night, and facilitate households’ factors of income generation. As mentioned above in the introduction, a household faces energy insecurity if its consumption is insufficient to meet the basic need of daily life in terms of requirements for cooking, lighting, washing/cleaning, warming/cooling the house. In Cambodia, the average annual electricity consumption is about 300 kWh/person (Shigeru and Han, 2019); however, its subsidy policy is applied to any household whose electricity consumption is less than 50 kWh/month or 600 kWh/year. With the average household size of 4.6 in Cambodia, this means that a person can receive electricity subsidy if she/he has annual electricity consumption less than 130 kWh, which is significantly below the average of electricity consumption in Cambodia. United Nations Development Programme (UNDP) (2010) uses a standard of electricity consumption within the range of 50–100 kWh per person per year to sustain life. Similarly, in the case of Sri Lanka, Tennakoon (n.d.) uses the quantitative threshold of per capita annual consumption of 120 kWh electricity for lighting and 35 kg of liquefied propane gas equivalent for cooking. However, some studies use an energy requirement per household cut-off at 2125 kWh by using Guatemala data for 1998–99 (Foster et al., 2000). The quantitative measures defining minimum energy needs are sometimes arbitrary because differences in cooking practices and heating/cooling by region, climate, and culture complicate the level of energy required (Shahidur et al., 2012). Some empirical works defined insufficient energy consumption as a condition where a household’s energy expense is more than 10% of its total income such that it will begin to impact general household welfare. The idea is that when households are forced to spend as much as 10% of income on energy, they are being deprived of other basic goods and services (Barnes, 2010). Based on the UK Department of Energy and Climate Change (DECC, 2015), a household will need to spend more than 10% of their income in order to 3
H. Phoumin, F. Kimura
Economic Modelling xxx (xxxx) xxx
3.3. Dataset, variables and multicollinearity test
acquire sufficient energy to have adequate standard of warmth. This study adopts both approaches of using the 10% threshold of household energy expenditure as an energy insecurity cut-off point, and the minimum 600 kWh/year household electricity consumption threshold. However, using the electricity approach as a cut-off point has a drawback as it is limited to only those houses that are connected to electrical grids.
The study uses the 2015 CSES dataset for analysis. CSES is a survey of households and their members on housing conditions, education, economic activities, production and income, level and structure of consumption, health, victimisation, vulnerability, energy source and consumption, etc. The CSES database at the National Institute of Statistics is open to external researchers for research and analysis; in 2015, it comprised 3840 households. Key variables of interest for the investigation include ‘energy insecurity’ and ‘welfare’, which are the ‘health outcome, school outcome, earning outcome, hours worked of children’ variables. The variable ‘energy insecurity’ is derived from the probability function based on several criteria as described in the methodology. Other household and community characteristics are described in Table 1. The above empirical model employs the multivariate regression models to investigate its hypotheses. The study also performed the multicollinearity test of the independent variables and it found that all the right-hand side variables are acceptable as the Variance Inflation Factor (VIF) is small well below 10 (see Table 2). However, to ensure the fitted coefficient and T-statistics are not inflated, the study used the robustness standard errors correction techniques to correct the estimation of T-statistics.
3.2. Empirical model In this study, we suspected that the energy insecurity of the households is likely to have direct impact on food consumption and on the education expenditure/consumption for children. Based on the aboveconceptualised energy insecurity, we can write the following: PðYi ¼ 1Þ ¼ IðYi ¼ E > 10%Þ if energy insecurity; PðYi ¼ 0Þ if otherwise where E is the share of energy expenditure to the total expenditure, I(.) is an indicator function that takes on a value of {1} if the bracketed expression is true, and a value of {0} if otherwise. Then I(.) equals the value of {1} and the household would be considered energy insecure. Thus, incidence of energy insecurity is:
4. Results and discussions
n 1 X Energy Insecurityi ¼ ½IðYi ¼ E > 10%Þ N i¼1
Based on the provided definition, it is estimated that 27% of the total households remain energy insecure. However, when broken down into on-grid and off-grid, 13% of the households connected to the grid still
A similar approach is applied to household energy insecurity using electricity consumption with the threshold of 50 kWh/month, or 600 kWh/year. Once the energy insecurity indicator has been determined, it is very important to investigate whether it has any impact on households’ welfare. The hypothesis is that household of energy insecurity is believed to substantially impact a households’ welfare including food consumption, education, and health of the individual in the household. To investigate energy insecurity on households’ welfare, two structural equations are constructed. The right-hand side of the equation in the first structural equation uses the independent variable ‘Energy Insecurity’ explicitly as it will affect households’ welfare. The second structural equation, the righthand side variable ‘Share of energy expenditure to total expenditure’, is used to investigate the magnitude of impact. Thus, the model specification can be written as: Welfarei ¼ α0 þ β1 Energy Insecurityi þ β2 Xi þ β3 Rurali þ Ui1
Table 1 Description of variables. Description of variables
Mean statistics
Log household food consumption
Logarithm of household food consumption in the past 12 months (KR) Logarithm of household education/ expenditure in the past 12 months (KR) Share of energy expenditure to total expenditure (Percentage)
15.99,678
Log household education consumption/ expenditure Share of energy expenditure to total expenditure Household size Log household’s income
(1)
Electricity access Without piped water access
Welfarei ¼ α0 þ β1 Share of Energy Expenditurei þ β2 Xi þ β3 Rurali þ Ui1 (2)
Without tubed well access
where ‘welfare’ is the dependent variable representing a household’s food consumption, education consumption/expenditure, child’s schooling outcome, and health outcomes. The proxy variable ‘schooling outcome’ is an index of Schooling Attainment Relative to Age (SAGE) variable (Han, 2015). The SAGE index is simply derived from the equation below:
Household head with no education Household head with primary education incomplete Household head with primary education complete Household head with upper secondary education
SAGE ¼
Name of variable
Years of schooling 100 Age Age of School Entry
Household head with university Electricity consumption per capita Female household head
The higher the index, the better the schooling outcome of children. The variable Xi is the set of exogenous variables representing the household’s characteristics such as household’s income, electricity access, electricity consumption per capita, education, and access to clean water. The variable Rurali is the community characteristic if the household is residing in the rural or otherwise.
Rural
6.129,855
Household size Logarithm of household’s income in the past 12 months ¼ 1 if household has electricity access; ¼ 0 for otherwise ¼ 1 if without piped water access; ¼ 0 for otherwise ¼ 1 if without tubed well access; ¼ 0 for otherwise ¼ 1 if household head with no education; ¼ 0 for otherwise ¼ 1 if household head with primary education incomplete; ¼ 0 for otherwise ¼ 1 if household head with primary education complete; ¼ 0 for otherwise
4.506,642 16.4995
¼ 1 if household head with upper secondary education; ¼ 0 for otherwise ¼ 1 if household head with university degree; ¼ 0 for otherwise Electricity consumption per capita (kWh/capita/year) ¼ 1 if female household head; ¼ 0 for otherwise ¼ 1 if rural; ¼ 0 for otherwise
.0255,275
KR ¼ Cambodian riel. Source: 2015 Cambodia Socio-Economic Survey. 4
13.39,273
.7272017 .0080,771 .2342366 .1802553 .4464704
.1875488
.041,417 252.5145 .756,968 .5965095
H. Phoumin, F. Kimura
Economic Modelling xxx (xxxx) xxx
Table 2 Multicollinearity test.
Table 4 Regression coefficient estimates of food consumption.
Variable
VIF
1/VIF
Household head with primary education complete Household head with no education Household head with primary education incomplete Electricity consumption per capita Rural Log household’s income Energy insecurity of household Household size Household head with university Electricity access Household head with upper secondary education Female household head Without tubed well access Without piped water access Mean VIF
2.82 2.39 2.12 1.59 1.53 1.44 1.34 1.34 1.34 1.32 1.19 1.13 1.12 1.01 1.55
0.354,838 0.417,612 0.471,108 0.629,142 0.655,284 0.692,652 0.746,116 0.746,855 0.748,552 0.756,013 0.840,303 0.888,105 0.888,915 0.991,721
Ongrid Offgrid Total
Total
914 87.30% 1898 68.00% 2812 73.27%
133 12.70% 893 32.00% 1026 26.73%
1047 100% 2791 100% 3838 100%
Coefficient
Coefficient
Energy insecurity of household Share of energy expenditure to total expenditure Household size
None
-.3871497*** (.0321,373) None
Without piped water access Without tubed well access Household head with no education Household head with primary education incomplete Household head with primary education complete Household head with upper secondary education Household head with university Electricity consumption per capita Female household head Rural Constant
-.0413,496*** (.0081,013)
Multiple regression model using Robust Standard Error Correctionb
.126,217*** (.0050,765) .0720,751*** (.0093,409) .0819,193*** (.0145,262) -.1786928*** (.0529,703) -.0935,054*** (.0133,904) -.1081735*** (.0220,311) -.0766,681*** (.0196,922)
.1115705*** (.0047,435) .1010312*** (.0095,474) .0762,213*** (.0150,818) -.2073098** (.0637,238) -.1066915*** (.0140,031) -.1493641*** (.023,567) -.1048406*** (.0207,686)
-.0120,718 (.0202,931)
-.0290,963 (.0223,665)
-.0264,382 (.0387,978)
-.0040,873 (.0420,527)
.0626,308** (.036,567) .0009257*** (.000142) .0864,377*** (.0139,653) -.137,467*** (.0144,018) 14.29,294*** (.1479394)
.1151599** (.0353,296) .0003998*** (.0000859) .0854,652*** (.0144,705) -.1628727*** (.0160,552) 13.84,663*** (.1443285)
Goodness of fit. Number of obs. ¼ 1155; F(14, 1140) ¼ 26.57; Prob > F ¼ 0.0000; Rsquared ¼ 0.2471; Root MSE ¼ 1.105. Note: (*) represents the significant level at 10%; (**) represents the significant level at 5%; and (***) represents the significant level at 1%. Number in bracket is the robust standard error. a The regression model includes the independent variable ‘share of energy expenditure to total expenditure’. b The regression model includes the independent variable ‘energy insecurity of the household’. Source: Author’s calculation.
The impact of a household’s income on welfare such as food consumption and education expenditure for children: The coefficient estimates of a household’s income show that food expenditure is likely to increase by about 7% for every doubling of a household’s income; similarly, education consumption and expenditure is expected to increase by about 17%. Thus, the impact of a household’s income has a direct benefit to welfare of the household. The impact of household head’s education on welfare such as food consumption and education’s expenditure for children: The coefficient estimates of household head’s completion of university has a positive impact on both food consumption (a 6% increase), and a positive impact on education expenditure/consumption (a 20% increase). The impact of other household and community characteristics on welfare such as food consumption and education’s expenditure for children: Among other household characteristics, its size has direct impact on food consumption and education of children. The characteristics of the community such as households living in rural area seems to have negative impact on food consumption as well as education
Table 3 Prevalence of household energy insecurity by off-grid and on-grid electricity. Households of energy insecurity (frequency and percentage)
Multiple regression model using Robust Standard Error Correctiona
Log household’s income Electricity access
face energy insecurity, while 32% of households living off-grid face energy insecurity (Table 3). To investigate the main hypothesis, the study regresses the households’ food consumption and education expenditure/consumption on the share of energy expenditure to the total expenditure and also on sets of household and community characteristics (Tables 4, 5 and 6). The result of coefficient estimates can be interpreted as follows: The impact of energy insecurity on welfare such as food consumption and education expenditure for the children: The coefficient estimates of share of energy expenditure to total expenditure’ in Table 3 showed that a household will likely reduce its food consumption by about 4.1% for every 1% increase in the share of energy expenditure to total expenditure. This means that every percentage point of energy expenditure will impact a household’s food consumption substantially. If the data allowed, this will further impact general household health, especially the children and elderly persons living in the household. Likewise, the coefficient estimates of ‘share of energy expenditure to total expenditure’ in Table 5 showed that households will likely reduce their education consumption/expenditure by about 5.8% for every 1% increase in the share of energy expenditure to total expenditure/consumption. This means that every percentage point of energy expenditure will impact the household’s education consumption/expenditure substantially. Again, energy insecurity likely reduces a household’s food consumption by about 38% compared to households that are energy secure. Likewise, energy insecure households are also likely to have reduced their education expenditure by 26% as compared to energy secure households. Thus, the impacts of energy insecurity are enormous as they affect direct food consumption and education expenditure, which are fundamental to support growth and human capital formation of children. It is believed that the less spending on education and less food consumption in households will further impact on children’s performance at school, especially their human capital formation. Therefore, the regression results in Table 6 shows that every percentage increase of ‘share of energy expenditure to the total expenditure’ is likely to reduce the children’s SAGE index by 47%.
Households of energy security (frequency and percentage)
Dependent variable ‘Log household food consumption’
Source: Author’s calculation. 5
H. Phoumin, F. Kimura
Economic Modelling xxx (xxxx) xxx
Table 5 Regression coefficient estimates of education consumption/expenditure. Dependent variable ‘Log household education consumption/ expenditure’
Energy insecurity of household Share of energy expenditure to total expenditure Household size Log household’s income Electricity access Without piped water access Without tubed well access Household head with no education Household head with primary education incomplete Household head with primary education complete Household head with upper secondary education Household head with university Electricity consumption per capita Female household head Rural Constant
Multiple regression model using Robust Standard Error Correctiona
Multiple regression model using Robust Standard Error Correctionb
Coefficient
Coefficient
None
-.2621564** (.124,238) None
-.0585,014*** (.0130,137)
Table 6 Regression coefficient estimates of schooling attainment relative to age (SAGE).
.0965,153*** (.0278,622) .1666177*** (.0484,125) .2905582*** (.1107727) .2213673 (.3583839) .0945,942 (.0918,691) -.6045892*** (.1477813) -.3900032*** (.0949,524)
.079,152** (.0266,817) .2035521*** (.0455,854) .2782467** (.1121011) .1756231 (.3712069) .0902,604 (.0931,144) -.6577931*** (.1492502) -.4435312*** (.0968,051)
-.0935,535 (.0977,389)
-.1159162 (.100,138)
-.0713,393 (.1437776)
-.0661,887 (.1488217)
.2170157* (.1482588) .0015,156*** (.0004294) .1336658 (.0829,068) -.5287331*** (.078,998) 10.29,859*** (.7414394)
.3188259** (.1445511) .0007297** (.0002679) -.1501823* (.0850,642) -.5313*** (.0800,466) 9.711,694*** (.712,983)
Dependent variable ‘SAGE’ - An index of schooling attainment relative to age
Multiple regression model using Robust Standard Error Correctiona
Multiple regression model using Robust Standard Error Correctionb
Coefficients
Coefficients
Energy insecurity
None
Share of energy expenditure to total expenditure Without piped water access Without tubed well access
-.4792831*** (.1270687)
2.06767** (1.049,695) None
Household head with no education Household head with primary education incomplete Household head with primary education complete Household head with upper secondary education Household head with university Electricity consumption per capita Female household head Rural Constant
1.160,382 (5.151,614) 2.150,683 (1.365,153) 13.5516*** (1.971,471) 9.285,921*** (1.562,142)
1.407,132 (5.177,143) 1.960,529 (1.36,468) 14.24,124*** (1.973,232) 9.720,728*** (1.563,496)
1.226,556 (1.690,172)
1.444,921 (1.70,026)
2.440,661 (2.553,496)
2.665,434) (2.58,741
11.25,608*** (2.553,982) .014,578*** (.0030,274) .3292066 (1.288,788) 4.979,571*** (1.212,425) 82.81,464*** (1.957,626)
12.80,953*** (2.523,761) .0067,002** (.0029,038) .3081961 (1.288,737) 5.220,719*** (1.227,291) 82.4632*** (1.983,757)
Goodness of fit. Number of obs. ¼ 3837; F(14, 3822) ¼ 208.74; Prob > F ¼ 0.0000; Rsquared ¼ 0.4795; Root MSE ¼ 0.3566. Note: (*) represents the significant level at 10%; (**) represents the significant level at 5%; and (***) represents the significant level at 1%. Number in bracket is the robust standard error. a The regression model includes the independent variable ‘share of energy expenditure to total expenditure’. b The regression model includes the independent variable ‘energy insecurity of the household’. Source: Author’s calculation.
Goodness of fit. Number of obs. ¼ 1155; F(14, 1140) ¼ 27.92; Prob > F ¼ 0.0000; Rsquared ¼ 0.2739; Root MSE ¼ 1.0851. Note: (*) represents the significant level at 10%; (**) represents the significant level at 5%; and (***) represents the significant level at 1%. Number in bracket is the robust standard error. a The regression model includes the independent variable ‘share of energy expenditure to total expenditure’. b The regression model includes the independent variable ‘energy insecurity of the household’. Source: Author’s calculation.
5. Conclusion and policy implications Energy security has been widely known from the supply side perspective where leaders try to secure energy supplies and also reduce the dependency on fossil fuel imports. Generally, leaders try to build an energy policy framework that shields countries from supply disruption risks. However, energy insecurity occurs when households are deprived of the energy needed to meet the basic minimum requirements for cooking, lighting, heating/cooling. Since the concept of energy insecurity fluctuates, the study adopted the share of energy consumption being 10% to the total expenditure as the threshold of household energy insecurity. With this definition, the national incidence of energy insecurity is estimated at about 27% of the total households in Cambodia. The study further separates the incidence of energy insecurity between off-grid and on-grid households. The result shows about 13% of the households connected to the grid still face energy insecurity; about 32% of households living off-grid face energy insecurity. The major findings about household food consumption show that it will likely decrease by about 4.1% for every 1% increase in the share of energy expenditure to total expenditure, with education consumption/ expenditure similarly being reduced by about 5.8%. Energy insecure households consume about 38% less than energy secure ones and will
expenditure. This result indicates that households spent less on food and education of children if they lived in rural or remote areas. The result also points to the fact that households living in rural areas are engaged in a subsistence economy where their labour is not counted and are making their own living by farming. The results also show that a household head having finished a university degree has substantial positive impacts on a household’s food consumption and education expenditure. Likewise, a household head with just a primary education or less has negative impacts on food consumption and education expenditure. This is self-evident because education is crucial to household achieving household income, and it further affects purchasing power. Finally, other community characteristics such as electricity access, and access to water sources such as piped or tubed well are keys for a positive impact on welfare of the households. The access to electricity and water sources represent the impact of infrastructure on welfare.
6
H. Phoumin, F. Kimura
Economic Modelling xxx (xxxx) xxx
likely spend 26% less on education. Thus, the impacts of energy insecurity are enormous. The study further found that every percentage point increase in the share of energy expenditure to the total expenditure is likely to reduce children’s school performance by 47%. For every doubling of a household’s income, the study found that it increases food expenditure by about 7%, with education consumption and expenditure similarly expected to increase by about 17%. Thus, the impact of household’s income has direct benefit to welfare of the household. The household-head’s completion of university has positive impact on both food consumption (a 6% increase), and education expenditure/consumption of children (a 20% increase). Among other household characteristics, its size has direct impact on food consumption and education expenditure, while residency in rural areas seems to have negative impact. The other community characteristics such as access to electricity and water sources represent the positive impact of infrastructure on household welfare. Thus, this study confirmed that energy insecurity has enormous negative impact on households and the human capital formation of the children. The above findings imply the following:
ASEAN Centre for Energy [ACE], 2019. Renewable energy: outcome-based Strategy. http://www.aseanenergy.org/programme-area/re/. (Accessed 8 September 2019). Barnes, D., 2010. The concept of energy poverty,’ energy for development and poverty reduction, 19 June. http://www.energyfordevelopment.com/2010/06/energy-pove rty.html. (Accessed 27 February 2019). Chen, Yin E., Fu, Qiang, Chang, Chun-Ping, 2019. International Sanctions’ Impact on Energy Efficiency in Target States, Economic Modelling. Economic Modelling, 23 July 2019. Dept for Energy and Climate Change (DECC) (2015), 2015. Annual Fuel Poverty Statistics Report. DECC, London. Electricity Authority of Cambodia [AEC], 2019. Electricity tariff- in Khmer language. https://www.eac.gov.kh/document/tariffdecidelist. (Accessed 10 September 2019). Foster, V., Tre, J.-P., Wodon, Q., 2000. Energy Prices, Energy Efficiency, and Fuel Poverty. World Bank, Washington, DC. Han, P., 2015. Energy Security and Sustained Growth: Analysis of Energy Outlook and Savings Potential in the EAS Region. Asian Development Bank Institute, Tokyo. https ://www.asiapathways-adbi.org/2015/11/energy-security-and-sustained-grow th-analysis-of-the-energy-outlook-and-savings-potential-in-the-eas-region/. (Accessed 18 February 2019). Han, P., Kimura, F., 2019. Cambodia’s energy poverty and its effects on social wellbeing: empirical evidence and policy implications. Energy Policy 132, 283–289. International Energy Agency (IEA), 2017. Southeast Asia Energy Outlook 2017: from Poverty to Prosperity. World Energy Outlook Special Report. IEA, Paris. Kutani, I., 2013. Study on the Development of an Energy Security Index and an Assessment of Energy Security Policy for East Asia Countries. Economic Research Institute for ASEAN and East Asia, Jakarta. http://www.eria.org/publications /study-on-the-development-of-an-energy-security-index-and-an-assessment-of-en ergy-security-for-east-asian-countries/. (Accessed 18 February 2019). Mingbo, Zheng, Chen, Yin E., Feng, Gen-Fu, Wen, Jun, Chang, Chun-Ping, 2018. Divorce and housing price in 31 provinces of China. Bulletin of Monetary Economics & Banking 21 (3), 173–188. Ministry of Mines and Energy of Cambodia (MME), 2016. Cambodia National Energy Statistics 2016. Economic Research Institute for ASEAN and East Asia, Jakarta. http: //www.eria.org/publications/cambodia-national-energy-statistics-2016/. (Accessed 19 February 2019). Ministry of Mines and Energy of Cambodia (MME), 2019. Cambodia Basic Energy Plan. Economic Research Institute for ASEAN and East Asia, Jakarta. http://www.eria. org/publications/cambodia-basic-energy-plan/. (Accessed 26 March 2019). Otaka, Y., Han, P., 2016. Study on the Strategic Usage of Coal in the EAS Region: A Technical Potential Map and Update of the First-Year Study. Economic Research Institute for ASEAN and East Asia, Jakarta. http://www.eria.org/publications/s tudy-on-the-strategic-usage-of-coal-in-the-eas-region-a-technical-potential-map-andupdate-of-the-first-year-study/. (Accessed 18 February 2019). Papada, L., Kaliampakos, D., 2016. Measuring energy poverty in Greece. Energy Policy 94, 157–165. Phnom Penh Post, 2019. Kingdom Lacks up to 400 MW in Available Electricity. https://www.phnompenhpost.com/national/kingdom-lacks-400mw-available-electri city. (Accessed 9 September 2019). Shahidur, K., Barnes, D., Samad, H., 2012. Are the energy poor also income poor? Evidence from India. Energy Policy 47, 1–12. Shigeru, K., Han, P., 2019. Energy Outlook and Energy Saving Potential in East Asia 2019. Economic Research Institute for ASEAN and East Asia, Jakarta. http://www.eria.org /publications/energy-outlook-and-energy-saving-potential-in-east-asia-2019/. (Accessed 26 March 2019). Cambodia Socio-Economic Survey (CSES), 2015. Cambodia Socio-Economic Survey. National Institute of Statistics, Cambodia. Ministry of Planning. The ASEAN Post, 2018. Southeast Asia’s Hope for Renewable Energy in the Year Ahead. https://theaseanpost.com/article/southeast-asias-hope-renewable-energy-year-ahea d. (Accessed 9 September 2019). United Nations, 2011. 2012 International Year for Sustainable Energy for All. htt p://www.un.org/en/events/sustainableenergyforall/. (Accessed 19 February 2019). United Nations Development Programme (UNDP), 2010. Energy for a Sustainable Future: the Secretary-General’s Advisory Group on Energy and Climate Change Summary Report and Recommendation. UNDP, New York. US Energy Information Administration (EIA), 2018. Residential energy consumption survey. https://www.eia.gov/consumption/residential/. (Accessed 22 February 2019). World Bank, 2016. Sustainable Development Goal on Energy (SDG7) and the World Bank Group. http://www.worldbank.org/en/topic/energy/brief/sustainable-develop ment-goal-on-energy-sdg7-and-the-world-bank-group. (Accessed 21 February 2019). World Bank, 2019a. More People Have Access to Electricity than Ever before, but World Falling Short of Sustainable Energy Goals. https://www.worldbank.org/en/news /press-release/2019/05/22/tracking-sdg7-the-energy-progress-report-2019. (Accessed 10 September 2019). World Bank, 2019b. Access to Electricity: % of Population. https://data.worldbank .org/indicator/EG.ELC.ACCS.ZS. (Accessed 21 February 2019).
o Welfare impacts: policymakers will need to identify households whose energy consumption is insufficient to meet basic needs. Once they are identified, the appropriate energy policy should target those vulnerable households to support them to have sufficient energy. Electricity access is the first priority, followed by a subsidy. In addition, the policy needs to address off-grid and on-grid families differently. o Infrastructure: Access to electricity, piped water systems, and other infrastructure are fundamental to increase households’ opportunities and income. Nonetheless, access to electricity is the first step for households to consume clean electricity in comparison to biomass or kerosene for heating and lighting. Thus, policymakers will need to find appropriate policies to ensure that households are connected to the electrical grid as fast as possible. While the development of the power grid and distribution will remain slow due to investment patterns, other measures, such as rooftop solar photovoltaics, solar farms, and stand-alone small generators will provide fast access to remote areas. Policies to support these distributed energy systems will need to be designed to support the investment in this area. o The government may need to review the current policy of subsidies for households whose electricity consumption is below 50 kWh/ month, as this threshold is too low to meet basic needs. The threshold could be increasing up to double in the first step and check how it impacts the welfare of the household. Thus, further study will be needed to see how much subsidy the government can afford, and to what extent the subsidy could have positive long-term impact on household welfare. Acknowledgements This study is carried out under academic project funding from Economic Research Institute for ASEAN and East Asia [ERIA]. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.econmod.2019.09.024. References ASEAN, 2007. Cebu declaration on East Asian energy security. Cebu, Philippines, 15 January. https://asean.org/?static_post¼cebu-declaration-on-east-asian-energy-secu rity-cebu-philippines-15-january-2007-2. (Accessed 18 February 2019).
7