World Energy Outlook Methodology for Energy Access Analysis

Transcription

1 1 Table of Contents Page 1 2 Summary Energy access methodology Defining modern energy access Electricity access Database Outlook for electricity access Reliance on traditional use of biomass Database Outlook for access to clean cooking facilities The Energy for All Case Achieving universal electricity access Achieving universal access to clean cooking facilities Energy Development Index Framing energy development Defining a new Energy Development Index Calculating the EDI Changes compared to previous EDI Application of the new EDI Areas for potential further development of the EDI... 13

2 2 Summary Modern energy services are crucial to human well-being and to a country s economic development. Access to modern energy is essential for the provision of clean water, sanitation and healthcare and for the provision of reliable and efficient lighting, heating, cooking, mechanical power, transport and telecommunications services. The World Energy Outlook (WEO) has for more than a decade devoted attention to the topic of energy access, informing the international community with key quantitative analyses, including energy access databases, projections and estimates of the investment needs and implications for global energy use and carbondioxide (CO 2 ) emissions of universal energy access. The latest contribution to the debate is the WEO-2012 chapter Measuring Progress towards Energy for All. This chapter has a special focus on the critical issue of how to track universal modern energy access and presents a new Energy Development Index. The first part of this methodology note discusses the WEO energy access databases, projections and investment model. The second part zooms in on the methodology for our Energy Development Index. Supplementary energy access materials are available on our website: There is no single internationally-accepted and internationally-adopted definition of modern energy access. In our analysis we define those lacking modern energy access as being those without access to electricity and those without access to clean cooking facilities. These are measured separately. We maintain databases on levels of national, urban and rural electrification rates and on the proportion of the population that rely on the traditional use of biomass for cooking, such as a three-stone fire. Both databases are regularly updated and form the baseline for WEO energy access projections to Due to differences in definitions and methodology from different sources, data quality may vary from country to country. WEO projections for access to electricity and to clean cooking facilities are based on separate econometric panel models that regress the electrification rates and rates of reliance on biomass respectively over many variables at a regional level. The resulting models relate access rates to per capita income, population growth, urbanisation, fuel prices, level of subsidies, technological advances, energy consumption and energy access programmes. We run these sub-models under the economy and population assumptions of the New Policies Scenario, the central scenario in WEO Investment requirements, fuel demand and CO 2 emissions are based on the regional additional power generation mix for electricity access, whereas for clean cooking a set of assumptions about clean cookstoves, LPG stoves and domestic biogas systems is used. The Energy Development Index (EDI) is a multi-dimensional indicator that tracks energy development countryby-country, distinguishing between developments at the household level and at the community level (Figure 1). In the former, it focuses on two key dimensions: access to electricity and access to clean cooking facilities. When looking at community level access, it considers modern energy use for public services (e.g. schools, hospitals and clinics, water and sanitation, street lighting) and energy for productive use, which deals with modern energy use as part of economic activity (e.g. agriculture and manufacturing). The choice of indicators used is constrained by the type of data related to energy access that is currently available. Figure 1 Components of the Energy Development Index * The geometric mean of the two variables is taken. ** Excludes electricity to avoid double counting. *** Includes industry, agriculture, services, transport and other non-specified energy use.

3 3 Energy access methodology The World Energy Outlook s activities in the domain of modern energy access consist both of analysis that is regularly updated and additional analysis on specific issues. Our general analysis consists of an update of our energy access databases, our energy access projections to 2030 and estimating the investment need and impacts on energy demand and CO 2 emissions. In this part of the methodology note we deal with the definition of energy access, data gathering for our energy access databases and the methodology for projecting future access rates. We also describe the methodology for our Energy for All Case, which is used to calculate additional investment needs and the additional impact on global carbon dioxide emissions from achieving universal modern energy access. 3.1 Defining modern energy access There is no universally-agreed and universally-adopted definition of modern energy access. For example, the UN Secretary General s Advisory Group on Energy and Climate (AGECC) defines energy access as access to a basic minimum threshold of modern energy services for both consumption and productive uses. Access to these modern energy services must be reliable and affordable, sustainable and where feasible, from low-ghgemitting energy sources. 1 GIZ recommends an orientation to specific service indicators for lighting, communication and food preparation and, for instance, for electricity access defines five service levels, corresponding to a certain package and kwh per capita consumption. 2 For our analysis, we define modern energy access as a household having reliable and affordable access to clean cooking facilities, a first connection to electricity and then an increasing level of electricity consumption over time. By defining access to modern energy services at the household level, it is recognised that some other categories are excluded, such as electricity access to businesses and public buildings that are crucial to economic and social development, i.e. schools and hospitals. Access to electricity involves more than a first supply connection to the household; our definition of access also involves consumption of a specified minimum level of electricity, the amount varies based on whether the household is in a rural or an urban area. The initial threshold level of electricity consumption for rural households is assumed to be 250 kilowatt-hours (kwh) per year and for urban households it is 500 kwh per year. The higher consumption assumed in urban areas reflects specific urban consumption patterns. Both are calculated based on an assumption of five people per household. In rural areas, this level of consumption could, for example, provide for the use of a floor fan, a mobile telephone and two compact fluorescent light bulbs for about five hours per day. In urban areas, consumption might also include an efficient refrigerator, a second mobile telephone per household and another appliance, such as a small television or a computer. The assumed threshold levels for electricity consumption are consistent with previous WEO analyses. However, we recognise that different levels are sometimes adopted in other published analysis. Sanchez, for example, assumes 120 kwh per person (600 kwh per household, assuming five people per household). 3 As another point of reference, the observed electricity consumption in India in 2009 was 96 kwh per person in rural areas and 288 kwh in urban areas on average over all people connected to electricity, implying a lower consumption for those that have been connected more recently. 4 Once initial connection to electricity has been achieved, the level of consumption is assumed to rise gradually over time, attaining the average regional consumption in the New Policies Scenario without universal access level after three years. This definition of electricity access includes an initial period of growing consumption as a deliberate attempt to reflect the fact that eradication of energy poverty is a long-term endeavour. In our analysis, the average level of electricity consumption per capita across all those households newly connected over the period is 750 kwh in Our definition of energy access also includes provision of cooking facilities which can be used without harm to the health of those in the household and which are more environmentally sustainable and energy efficient than the average biomass cookstove currently used in developing countries. This definition refers primarily to biogas systems, liquefied petroleum gas (LPG) stoves and advanced biomass cookstoves that have considerably 1 The Secretary General s Advisory Group on Energy and Climate Change (AGECC), Energy for a Sustainable Future: Summary report and recommendations, United Nations, 28 April 2010, New York. 2 See: and accessed November Sanchez, T. (2010), The Hidden Energy Crisis: How policies are failing the world s poor, Practical Action Publishing, London 4 Government of India (2011), Key Indicators of Household Consumer Expenditure in India, , Ministry of Statistics and programme Implementation, Government of India, New Delhi

4 lower emissions and higher efficiencies than traditional three-stone fires for cooking. In our analysis, we assume that LPG stoves and advanced biomass cookstoves require replacement every five years, while a biogas digester is assumed to last 20 years. Related infrastructure, distribution and fuel costs are not included in our estimates of investment costs. Our definition is intended to be supportive of the objective to conduct forward-looking projections, but data availability means that it is not viable to apply it to our estimates of the number of people that do not currently have access to modern energy services. This definition cannot be applied to the measurement of actual data simply because the level of data required does not exist in a large number of cases. As a result, our energy access databases focus on a simpler binary measure of those that do not have access to electricity and those that rely on the traditional use of biomass for cooking. This is disaggregated (either with data or estimation) between those in urban and rural areas within a given country. This level of reporting is far from optimal but is driven by severe data limitations. The number of people relying on traditional biomass for cooking is used as a proxy for measuring those that do not having access to clean cooking facilities (again, due to data constraints). We define the term biomass broadly to cover the various resources like wood, charcoal, tree leaves, crop residues and animal dung. The traditional use of biomass refers to the basic technology used, such as a three-stone fire, traditional mud stoves or metal, cement and pottery or brick stoves, with no operating chimneys or hoods. As a consequence of the pollutants emitted by these inefficient devices, pollution levels inside households cooking with biomass are often many times higher than typical outdoor levels, leading to about 2 million premature death each year according to the World Health Organization. 5 It is noted that there are some forms of cookstove that use traditional biomass in a safer, less-polluting way than some liquid fuels but, while the numbers are believed to be relatively small, it is not currently possible to distinguish these in the data. 3.2 Electricity access Database The general paucity of data on electricity access means that it must be gathered through a combination of sources, including: IEA energy statistics; a network of contacts spanning governments, multilateral development banks and country-level representatives of various international organisations; and, other publicly available statistics, including US Agency for International Development (USAID) supported DHS survey data, the World Bank s Living Standards Measurement Surveys (LSMS), the UN Economic Commission for Latin America and the Caribbean s (ECLAC) statistical publications, and data from national statistics agencies. In the small number of cases where no data could be provided through these channels other sources were used. For many countries, data on the urban and rural breakdown was collected, but if not available an estimate was made on the basis of pre-existing data or a comparison to the average correlation between urban and national electrification rates. Often only the percentage of households with a connection is known and assumptions about an average household size are used to determine access rates as a percentage of the population. To estimate the number of people without access, population data comes from OECD statistics in conjunction with the United Nations Population Division reports World Urbanization Prospects: the 2009 Revision Population Database, and World Population Prospects: the 2010 Revision. Electricity access data is adjusted to be consistent with demographic patterns of urban and rural population. Due to differences in definitions and methodology from different sources, data quality may vary from country to country. Where country data appeared contradictory, outdated or unreliable, the IEA Secretariat made estimates based on cross-country comparisons and earlier surveys Outlook for electricity access The electricity access database provides valuable information about the current electrification rates in a large number of countries. In order to provide an outlook for electricity access in the next decades, a model able to generate projections of electrification rates by region has been developed. The projections are based on an econometric panel model that regresses historic electrification rates of different countries over many variables, to test their level of significance. Variables that were determined statistically significant and consequently included in the equations are: 5 See:

5 per capita income demographic growth urbanisation level fuel prices subsidies to electricity consumption technological advances electricity consumption electrification programmes The model was developed within the framework of the WEO New Policies Scenario, using the same economic and population assumptions. The New Policies Scenario takes account of the broad policy commitments and plans that have been announced by countries around the world, to tackle either environmental or energysecurity concerns as well as plans to phase out fossil-energy subsidies. However, we assume that no additional policies to expand energy access, over and above those in place today, are enacted over the projection period. 3.3 Reliance on traditional use of biomass Database Our database on the traditional use of biomass for cooking makes use of the World Health Organization s (WHO) Global Health Observatory estimates of reliance on solid fuels. 6 In line with the level of data available from most sources, it focuses on the population where solid fuels are the primary fuel for cooking. We make an adjustment to subtract coal use from WHO totals, based on the most recent national data. 7 At the time of publication, a significant difference between our own database and that of the WHO is the data reported for India. This is because we make use of the results from the 2011 Census, which has yet to be included in the WHO dataset. Precise numbers on stove type or quality and on secondary sources of fuel for cooking are not available for most countries. Once again, we combine this data with population estimates from OECD statistics in conjunction with the United Nations Population Division reports World Urbanization Prospects: the 2009 Revision Population Database, and World Population Prospects: the 2010 Revision. Reliance on biomass data was adjusted to be consistent with demographic patterns of urban and rural population Outlook for access to clean cooking facilities To provide an outlook for the number of people relying on the traditional use of biomass in the next decades, a regional model was developed under the New Policies Scenario assumptions. Reliance on biomass rates of different countries is econometrically projected using many variables to assess their level of significance. Variables that were determined statistically significant and consequently included in the equations are: per capita income demographic growth urbanisation level level of prices of alternative modern fuels subsidies to alternative modern fuel consumption technological advances clean cooking programmes As for the electrification model, we assume that there are no additional policies implemented to reduce reliance of households on traditional biomass for cooking, over and above those in place today. 3.4 The Energy for All Case In our New Policies Scenario, we project that almost 1 billion people will lack access to electricity and 2.6 billion people will still be without clean cooking facilities in In our Energy for All Case, we examine the trajectory that would be required to achieve the goal of universal access to electricity and clean cooking 6 For more information, see 7 There are around 400 million people that rely on coal for cooking and heating purposes, which causes air pollution and has serious potential health implications when used in traditional stoves. These people are mainly in China, but there are also significant numbers in South Africa and India.

6 facilities by 2030 and what the implications would be of doing so. 8 The Energy for All Case quantifies the number of people that need to be provided access to modern energy services and the scale of the investments required by 2030 in order to achieve universal access to modern energy services Achieving universal electricity access The Energy for All Case is based on the assumption that new policies are introduced that result in a progressive increase in electrification rates to 100% of the world s population by Achieving universal access to electricity by 2030 would result in higher global energy demand than projected in the New Policies Scenario. It would also have implications for energy investment and for energy-related CO 2 emissions. The Energy for All Case seeks to quantify these increments and is calculated in four main steps. 1. Determine the additional number of people gaining access to electricity every year to reach universal electrification in 2030 and their consumption pattern Those gaining access in the New Policies Scenario are taken as the foundation upon which a gradually increasing number of additional people are estimated to need to receive electricity access in order to attain universal access by This therefore results in a projection of the additional number of people by region gaining access to electricity each year compared to the New Policies Scenario. The Energy for All Case assumes a basic level of electricity consumption for each person gaining access, as well as its evolution over time, in order to estimate the additional demand compared to the New Policies Scenario projection to It is assumed that each person gaining access is at first going to use electricity only as a substitute for the traditional fuels used to cover basic needs (e.g. candles, liquefied petroleum gas [LPG], kerosene). To assess the extent of the additional generating capacity required to achieve universal access, we have made assumptions about minimum levels of consumption at both the rural and urban level (see definition section above). Although the starting consumption is the same across regions, there is a regional difference in the catching-up process, depending on the average consumption of each region under the New Policies Scenario and the rural and urban disparity. 9 By 2030 the average per-capita electricity demand is 750 kwh. 2. Determine the additional capacity needed for power generation To estimate the need for additional generation capacity needed (in GW), we match the additional demand from people getting access in the Energy for All Case to the existing residential demand, total electricity generation and generation capacity in the New Policies Scenario. We take into account losses and own electricity use by the power generation sector for grid-supply. For urban-area electricity demand, the less costly choice is electricity grid extension, thus the model assumes that generation in urban areas is made entirely through grid options. In rural areas, options to increase electrification include extension of existing grids, creation of mini-grids and isolated off-grid generation. 10 That part of the rural area around onethird of total rural demand closest to urban areas and/or likely to become more densely populated by 2030 is also projected to be supplied through the grid, as this will be the most economic option. The remaining rural generation is off-grid, divided between mini-grids and isolated off-grid generation, including PV, mini-hydro, biomass, wind, diesel and geothermal. Mini-grids, defined as village- and districtlevel networks with loads of up to 500 kw, will constitute the bulk (65%) of off-grid generation, while isolated off-grid options will provide electricity to the remotest populations. The Energy for All Case assumes a breakdown by power generation options similar to the latest additions per region. This enables us to approximate the real generation mix for additional electricity access. For mini- and off-grid solutions we correct for typical distributed generation mix options found in richer areas by assuming a higher mix of diesel-generated or diesel-pv generated solutions. 3. Quantify the investment need for additional power generation and transmission and distribution To evaluate the cost associated with generation, the Energy for All Case feeds the added required generation (in GW added capacity) between 2011 and 2030 into the power generation module of the 8 In WEO-2010 the Energy for All Case was referred to as the Universal Modern Energy Access Case (UMEAC). 9 By consumption we mean electricity demand from the residential sector, which differs from demand for generation by excluding transmission and distribution losses and own use of the power generation sector. 10 Mini-grids provide centralised generation at a local level. They operate at a village or district network level, with loads of up to 500kW. Isolated off-grid solutions include small capacity systems, such as solar home systems, micro-hydro systems, wind home systems and biogas digester systems.

7 WEM 11 and the transmission and distribution model. Costs for off-grid generation are adjusted on the base of available costs for the various options (mini-grid and isolated off-grid) by region and country. 4. Calculate the implication of the additional electricity generation on energy-related CO 2 emissions The Energy for All Case quantifies the additional generation needed to have 100% electrification by Generating this additional electricity would impact on energy-related CO 2 emissions. The model assumes that the electricity is generated using the fuel mix calculated in the manner described above, using the mix of added capacity in the New Policies Scenario for the country or region in question. Additional CO 2 emissions are then calculated using the average CO 2 content of each power generation option by fuel Achieving universal access to clean cooking facilities In the New Policies Scenario, 2.6 billion people still rely on the traditional use of biomass as their primary source for cooking in 2030, with serious consequences for their health. The Energy for All Case shows a possible path towards universal access to clean cooking facilities (clean cooking fuels and stoves, advanced biomass cookstoves and biogas systems) by This is done in three steps. 1. Determine the additional number of people gaining access by technological solution, to reach universal access to clean cooking facilities in 2030 LPG stoves are judged to be more likely to penetrate in urban zones, where infrastructure, distribution and fuel costs can benefit from economies of scale and consumers have a relatively higher ability to pay. Thus LPG stoves are assumed to provide clean cooking services for all urban zones still relying on the traditional use of biomass but for only 30% of rural households. The large majority of rural households (55%) are assumed to be provided with advanced biomass cookstoves, and the remaining 15% with biogas home systems. 12 Those global targets are then reflected in regional allocations of the various options (LPG, biomass cookstoves and biogas systems) that are derived from assumptions regarding the most likely technology solution in each region, given resource availability and government policies and measures. 2. Quantify the investment costs Once the number of people that need each clean cooking facility option by country is determined (together with the rural/urban split), investment costs are calculated based on the unit cost of the different devices. Advanced biomass cookstoves, with emissions and efficiencies similar to those of LPG stoves, are assumed to cost $50. The assumed cost of an average-sized biogas digester varies by region. Based on 2010 data provided by SNV, the cost is $437 for India, $473 in China, $660 in Indonesia, $526 in other developing Asia, $702 in Latin America and $924 in sub-saharan Africa. An LPG stove and canister is assumed to cost $60. Infrastructure, distribution and fuel costs are not included in the investment costs. We assume that LPG stoves and advanced biomass cookstoves require replacement every five years, but only the cost of the first stove and half of the cost of the second stove is included in our investment projections. This is intended to reflect a path towards such investment becoming self-sustaining. 3. Calculate the related additional energy consumption and CO 2 emissions Expanding household access to modern fuels would inevitably increase global demand for these fuels, notably oil. Using World Heath Organisation (WHO) data for developing country households currently using LPG, we have estimated an average LPG consumption of 22 kg per person per year. The additional oil demand associated with access to LPG in the Energy for All Case is calculated based on the average per capita consumption (converted into mb/d of LPG) and the estimated number of people who switch to LPG stoves by The impact on CO 2 emissions of switching to advanced biomass technologies or LPG is very difficult to quantify because of the diversity of factors involved, including the particular fuels, the types of stoves and whether the biomass used is replaced by new planting and if a sustainable forestry management programme is in place. 13 But it is widely accepted that improved stoves and greater conversion efficiency would result in emissions reductions. As a worst-case approach we calculate the CO 2 emissions from LPG consumption of people gaining access to this solution for access to clean cooking. 11 The IEA s World Energy Model (WEM) is the underlying tool for the World Energy Outlook analysis. The WEM is a partial equilibrium model designed to replicate how energy markets function and provides medium to long-term energy projections. Documentation describing the overall approach and features of WEM may be found on 12 These are simple small-scale biogas digesters. 13 It is recognised increasingly that black carbon emissions from burning traditional biomass contribute to global warming.

8 4 Energy Development Index 14 Tracking progress through effective energy development indicators is a critical energy access issue. Since 2004, we have presented an Energy Development Index (EDI), which is devised as a composite measure of a country s progress in transitioning to modern fuels and modern energy services, as a means to help better understand the role that energy plays in human development. For WEO-2012 we have built on previous work, re-examining what variables have strong explanatory value for energy access, as well as the coverage, frequency and reliability of datasets relating to these variables. We have constructed a new EDI, which includes important additional indicators such as productive use of energy presented in the form of country level results. We have collected a substantial amount of data, allowing us to present EDI results for 80 countries and to compare EDI results over time. Such an indicator can help ensure that policy and financing commitments achieve maximum impact. 4.1 Framing energy development Energy development indicators should seek to quantify availability of modern fuels, the consumption of these fuels and how this consumption contributes to basic needs and human development. However, following this path from supply through to consumption and then impacts is fraught with challenges. Instead, we take a snapshot at a country level, segregating broadly between energy development at the household level and at the community level (Figure 2). In the former, we focus on two key dimensions (as reflected earlier in this chapter), that of access to electricity and access to clean cooking facilities. Access to heating is an important additional variable that is sometimes mentioned in this context, but is often either excluded due the lack of available data or dismissed as it is strongly related to cooking (as the same means may be used for both). When looking at community-level modern energy access, the categories are broader by necessity. In the case of public services, the focus is on examining the use of modern energy in schools, hospitals and clinics, water and sanitation, street lighting and other communal institutions. In the case of productive use, the focus is often on modern energy use for activities such as agriculture (ploughing, irrigation and food processing), and micro-small enterprises, such as milling, textiles etc. The issue of modern energy for productive uses is not a new one but is rightly receiving increased recognition. An additional aspect of modern energy use captured, to an extent, within productive use is that of transport. It is valuable to capture this variable as, particularly in earlier economic development, a significant share of its energy consumption is directly or indirectly for productive economic purposes. Figure 2 Energy development framework Notes: Household does not distinguish energy use as part of a micro-enterprise conducted within the home (which existing energy access data is often unable to identify). While very different in nature, energy for public services and for economic/productive purposes are grouped here under the community heading. Within these broad categories, access to modern fuel and the appliances to utilise it are considered together i.e. a person has adequate access only if they have access to both. However, it is recognised that, in respect to both access to energy and access to appliances, there is also a progression. For instance, in the case of electricity, the first move might be from candles and batteries to solar lanterns, solar home systems or, 14 This analysis benefitted from a roundtable meeting held by the IEA in Paris on 25 May 2012.

9 possibly, a mini-grid. Similarly, first access is likely to involve only a small number of basic appliances, with greater diversity coming later. In addition, there are a number of issues that are sometimes referred to generically as quality of supply. For any energy supply to provide a genuine opportunity to use modern energy services there needs to be a technical possibility to use it (availability), a price that is not prohibitive (affordability), sufficient supply (adequacy) and a supply that is easy to use (and pay for), including being located nearby, available at desired hours of the day and safe to use (convenience). Importantly, the supply must be of the right quality (e.g. voltage level) and be usable for most of the time (reliability). 15 At a more sophisticated level, it is also recognised that it may be desirable to track the quality of policies, regulations and institutions involved and, certainly, whether there is sufficient funding to support realisation of the objectives. 4.2 Defining a new Energy Development Index The EDI has as its primary focus to inform international policy makers on a regular basis on whether broadenergy development is taking place. While modern energy access can be considered from the perspective of multiple energy users, sources, uses etc, our ambition for the EDI continues to be to develop an annual multidimensional indicator that tracks energy development at a country level. Strengths of the EDI include its relatively broad coverage (i.e. large number of countries), the fact that it is updated annually and that it seeks to measure several elements of modern energy development individually and then as a composite index. However, it is recognised that the EDI could potentially be strengthened with regards to productive use, energy for public services and quality of supply; even if data for the ideal variables are not available. Focusing on the dimensions of modern energy access set out above, an examination of the options reveals that, in theory, there are a number of potential variables that can be monitored as a means of helping to assess energy development (Table 1). However, a key issue that continues to rule out many options is the availability of regular, reliable and robust data. This is because data on many possible variables is only collected as part of household or business surveys, which are often only conducted on an infrequent basis in many countries. While this persists, some compromise variables, with a good degree of explanatory value, are required. For this, we mainly rely on energy balance data from countries and, where possible and beneficial to do so, we use multiple data sources to measure or verify them. Our assessment of the strengths and weaknesses of different possible indicators leads us to select per capita public services electricity consumption as our public services variable in the EDI and, as a productive use variable, the share of industry, agriculture, services, transport and other sectors in total final energy consumption for a country. While there appear to be more appropriate measures to assess quality of electricity supply, the quality of data is currently weak and so excluded from the new EDI. Mapping our productive use variable against the UN Human Development Index 16 (HDI) reveals a relationship that appears to be relatively linear (Figure 4). There is a notable clustering of those that score relatively highly on the HDI and those that score higher on our productive use variable. Additional to the aspects included in Table 1 and Figure 3, there are also variables that could be measured from a project financing point of view with the purpose to link project activities to impacts on use of modern energy services. Other aspects that could be taken into account are whether the institutional and other (e.g. infrastructural) conditions are in place that are beneficial for increasing access to modern energy services. This requires monitoring of policy and infrastructure development. At the global level, the world might first be interested whether countries have targets and monitoring in place, and whether there is sufficient financing of energy access (by different sectors, which is a possible but demanding task). 17,18 15 We follow the proposed characterisation proposed by Pachauri in Reaching an international consensus on defining modern energy access, Current Opinion in Environmental Sustainability, Issue 3, Elsevier, Amsterdam, pp , The Human Development Index is a summary measure of human development. It measures the average achievements in a country in three basic dimensions of human development: a long and healthy life (health), access to knowledge (education) and a decent standard of living (income). 17 See the analysis presented in the WEO 2011 special excerpt "Energy for all - Financing access for the poor". 18 For an example of tracking energy access commitments of countries, see the UN Sustainable Energy for All website:

10 Table 1 Sample of possible indicators for measuring modern energy access variables Dimension Clean cooking Electricity access Energy for public services Productive use Reliability of electricity supply Potential indicator % of households having an efficient stove that meets minimum requirements for indoor air quality % of households that cook with modern fuels % of income or expenditure spent on cooking fuels % of households with access that meets minimum quality requirements Number of household electricity connections % of hospitals and schools with minimum level of electricity supply and % of villages with street lighting % of businesses with equipment common for their sector that consumes modern energy % of businesses with minimum quality of electricity supply % of farms with minimum availability of advanced mechanical power Average number of power outages per year and their average duration. Expenditure (as % of household or company income) on back-up electricity generation Strengths Directly measures the main issue relating to clean cooking facilities indoor air quality Close to a direct measurement. Does not require an inventory of stove types. It captures affordability and is an indication for how price changes will affect access Measures multiple aspects of household electricity access situation Numbers can be updated annually, and help illustrate availability Data could be gathered by relevant government ministries Ownership of modern equipment is a sign of using modern energy services Potentially available from utilities or electricity regulator Measures productive use of modern energy in an important sector Potentially available from utilities or electricity regulator This will have a strong correlation with the reliability of electricity supply Weaknesses Data improving but still not sufficient. Requires regular surveys adopting common standard for minimum acceptable air quality. Data improving but still not sufficient. Does not reflect use of improved cookstoves with traditional fuels. It does not capture gathering fuel wood. Requires regular expenditure surveying Data severely lacking and a common standard required for minimum quality Several cases were reported where connections do not equal effective electricity access. Also excludes illegal connections. Data not commonly collected at present. No common definition of minimum levels. Such detailed data is uncommon, as is it normally only collected in business surveys Minimum requirements differ between businesses and not reported at present. Not commonly reported at present and would need a definition for minimum availability. Not commonly reported at present It requires regular household expenditure and enterprise surveys, also rurally and in slums

11 Human Development Index World Energy Outlook Methodology for Energy Access Analysis Figure 4 HDI and new productive use variable for the EDI (2010) Productive use indicator 4.3 Calculating the EDI While the results for individual indicators tell a story in themselves for a country, it is the four indicators taken together that form the EDI, providing an overview of its energy development (Figure 6). The access to electricity indicator is based on two variables: The first one is the electricity access rate. This is the percentage of population that has access to electricity, mainly through a grid connection or an equivalent off-grid electricity supply. Data comes from IEA s electricity access database (see part 1 of this methodology note). The second variable is the per capita residential electricity consumption. For this variable, data comes from IEA s country energy balances and for some countries from the United Nations Statistics Division (UNSD). In some cases, residential electricity use is not reported individually, but together with other sectors, such as services and agriculture. In that case, additional assumptions have been made about the split of electricity use among these sectors, usually taking the regional average. These additional assumptions are indicated in the EDI database. To calculate the overall access to electricity, we take the geometric mean of two variables as an indicator. Figure 6 Components of the Energy Development Index * The geometric mean of the two variables is taken. ** Excludes electricity to avoid double counting. *** Includes industry, agriculture, services, transport and other non-specified energy use. For the access to clean cooking facilities indicator we use the share of modern energy use in the total final consumption by the residential sector. Modern fuels in this sense mean all fuels except solid fuels 19. We exclude electricity use, because it is already part of the household access to electricity indicator and there is no data yet to split electricity use for cooking out of the total residential electricity use at a country level for most countries in the EDI. Data comes from IEA s country energy balances and for some countries from the United Nations Statistics Division (UNSD). Access to energy for public services is represented by per-capita public services electricity consumption, for which we use an approximation. We first estimate the pubic sector s share in the services sector, and then 19 Modern fuels are: oil products, gas, renewables (own supply) and heat; the first part of the energy access methodology note deals with fuel use and access to clean cooking facilities.

12 multiply this with the services sector electricity consumption to then calculate the per-capita public services electricity consumption. To estimate the share of the public sector in the services sector we take the governmental final consumption 20 (percentage of GDP) as a fraction of the share of total services value added in GDP (World Bank data). For some countries these were not available and additional assumptions have been made. Where this was the case, this has been indicated in our EDI database. Also, in case services electricity consumption was not reported individually, additional assumptions were made. These additional assumptions take into account that some public services, like electricity use (mostly street lighting), is registered in the nonspecified/other sector s energy consumption. Data comes from IEA s country energy balances and for some countries from the United Nations Statistics Division (UNSD). For access to energy for productive use we calculate the share of economic purposes in the total final consumption (TFC). TFC includes the sectors: industry, transport, services, agriculture/forestry and fishery, and excludes own use by the energy sector and residential energy use. Data comes from IEA s country energy balances and for some countries from the United Nations Statistics Division (UNSD). To calculate the EDI, the values for the variables are first normalised. To do this we use fixed values, sometimes being a cut-off value based on a comparison of overall values and comparison to OECD countries (Table 2). We use fixed values as this supports year-to-year comparison of EDI scores. For per-capita electricity consumption in the residential sector we use a cut-off value of toe per-capita per year (1 163 kwh per capita). This is close to lower levels of per-capita electricity consumption in OECD countries in 2010 (Italy toe/capita, Korea toe/capita). For per-capita public services electricity consumption we use a cut-off value of toe. This is comparable to the highest scores of countries in our list (South Africa toe/capita, Malaysia toe/capita) and below that of lower OECD country scores (Hungary toe/capita, Poland toe/capita). Finally, for the productive use indicator we choose 90% as the cut-off value, as the highest developing economies can be in the range of 75-90%. Table 2 Variables and normalization values for calculating the 2012 Energy Development Index Indicator Minimum value Maximum value Share of population with access to electricity 0% 100% Per-capita electricity consumption in the residential sector (toe) Share of modern fuels in total final residential sector consumption 0% 100% Per-capita public services electricity consumption (toe) Share of industry, agriculture, services, transport and other sectors in total final energy consumption 0% 90% Once normalised, we take an average across the two household-level indicators and a separate average across the two community-level indicators. We then average the result of these two scores to reach a final EDI score. This process means that we apply equal weighting to each indicator in an effort to leave our methodology neutral to any judgement of whether individual indicators are more or less important than others. It is recognised that some implicit weighting takes place when choosing the way an indicator is calculated and normalised. For instance, an increase in public services electricity consumption is seen only in later stages of development, whereas the increase of productive use in total final consumption is observed to be a more linear process. This means that the productive use indicator is likely to be higher earlier than the public services electricity consumption. We have no basis on which to ground any adjustment to these trends and simply note this as an observation. We have tried to choose our cut-off values at such a level that they do not dwarf progress for countries at low levels, nor inflate the scores of high scoring countries. Overall scores and individual scores are published on our online EDI database at Changes compared to previous EDI There have been three main changes compared to our previous edition of the EDI. The first is a different choice of variables (see WEO-2010 for the four indicators that previously combined to form the EDI). The variables that have been added are the indicators for energy use for public services and for productive use. These have been added so as to capture these broader aspects of energy development more explicitly. A 20 This includes governmental consumption of products, but also salaries of people working in public services and can therefore be seen as a useful proxy for the size of government in the services sector.

13 second change is the restructuring of the EDI into a household-level element and a community-level element. This is not only a presentational issue, but also reflects a new way of calculating the household electricity access indicator. Whereas the topic of household electricity access previously was represented by two out of four variables, it now has one indicator that is equally valued as the access to clean cooking facilities indicator. The third change is the way in which normalisation takes place. We now normalise against fixed values, which is motivated by a desire to be able to compare individual country s scores over time (rather than just their ranking). 4.5 Application of the new EDI This EDI, presented by country, but also split by indicator and shown over time, should be a valuable aid to a range of decision makers in tracking progress in important elements of a country s energy development. The EDI will also become more valuable as data quality improves, reinforcing the need for initiatives, such as the UN Sustainable Energy for All initiative, to emphasise and support efforts to strengthen capacity in this area. However, the EDI should still be seen as part of a broader suite of indicators that might also provide coverage at project/programme, sub-national and regional levels. We will continue to update the EDI on a regular basis, reflecting the latest available data, and, whenever possible, seeking to expand our country coverage. We will also continue to review the range of data sets available and, in light of developments, consider if our EDI methodology can be strengthened further, to provide an even better measure of a country s energy development. 4.6 Areas for potential further development of the EDI We recognise that there is still potential room for further improvement of the EDI. The standardised definition and measurement of country data would help further improve the EDI. Stronger indicators for public services and productive use, such as discussed in Table 1, would be desirable if the data were available. Another desirable improvement would be the inclusion of a stronger indicator for quality of supply or efficiency of energy use. Other factors include the affordability of modern energy 21, how efficiently it is consumed and the level of consumption by small and medium-size enterprises 22 given their crucial role in employment and economic growth. An important issue for some countries or institutions might be the development of a low carbon EDI, in which the variables were specifically related to measurement of the role of renewables and low-carbon technologies in the energy system. This might be driven by a concern that energy development that relies on fossil fuels, even to a relatively limited extent, could result in the lock-in of these technologies, and their associated emissions, for decades to come. Examining either the level of electricity generated from renewables (or fossil fuels) as a proportion of overall electricity would be one obvious indicator. If this were unavailable, the proportion of renewable energy (or fossil fuels), in a country s energy mix could be a fall-back measure. The number of recorded sales/installations of solar home systems or other types of renewable technologies (or of various energy-efficient technologies) could also be measured. For cooking, it might be appropriate to examine not only the type of fuel used but also the efficiency of stoves that are commonly sold, as this will be an important determinant of fuel demand. In terms of environmental or impact indicators, the level of local pollution linked to burning hydrocarbons could and should be monitored, as well as the number of reported illnesses related to local pollution. In terms of public services, one could measure the number of public health centres and the percentage of schools or training centres that provide their services on the basis of renewable energy technologies. 21 See Winkler, H., Simoes, A. F., La Rovere, E. L., Atiq Rahman, M. A., & Mwakasonda, S. (2011). Access and Affordability of Electricity in Developing Countries. World Development, Vol.39, See Kooijman-Van Dijk, A. L., & Clancy, J. (2010). Impacts of Electricity Access to Rural Enterprises in Bolivia, Tanzania and Vietnam. Energy for Sustainable Development, 14 21