Energy Poverty and Income Poverty: How Do They Differ?

by Doug Barnes, Shahid Khandker and Hussain Samad

There is a continuing discussion over what constitutes energy poverty with several approaches being used to define it.  But as yet, no consensus has emerged for measuring and monitoring energy poverty and explaining why and how it differs from income poverty.  After all, income poverty is a standard measure, so if the two are highly related it would not be worth the effort to develop a unique measure for energy poverty.  In other words, why create a separate indicator of energy poverty because in the end it would just be a reflection of income poverty.  Everyone seems to know about energy poverty,but truly defining and measuring it gets a bit complicated. 

About 10 years ago energy poverty was thought to be related mainly to lack of access to electricity.  More recently the United Nations and Department of International Development of Great Britain (DFID) have broadened definitions of energy poverty to multiple indicators using somewhat arbitrary weights.  International Energy Agency (IEA) has never actually defined energy poverty (except that it is related to lack of access to modern energy), but advocated that better ways of using biomass energy for cooking should be an important policy for household energy.  Also, most international organizations measure energy poverty indicators as outputs (e.g., lack of electricity connections) rather than outcomes (e.g., welfare gains from electricity consumption).  Thus, unlike income poverty—which is usually based on minimum consumption of food and non-food items necessary to sustain a livelihood—energy poverty lacks a well-established theory based on energy demand to establish a relevant poverty line. 

In several recent papers the authors of this post have taken a different approach, focusing on energy demand in order to define energy poverty.  Like income poverty, energy poverty may be defined by the minimum energy consumption needed to sustain lives.  This approach defines an energy poverty line as a threshold of energy consumption needed to sustain life.  Similar to the concept of income poverty, we reasoned that there had to be a point at which energy is essential for living.  After all, people have to cook their food; in cold climates they must heat their homes; and they generally need a minimum level of light in the evening for basic tasks (sometimes including eating).  In theory this is all well and good, but the question remained how to measure that threshold. 

Figure 1:  Energy End Use Energy Consumption by Income Class,
Bangladesh and India

Therefore, we tested this demand approach using rich household survey data sets from Bangladesh and India.  We found that although energy consumption rises with income, this increase is not uniform (Figure 1).  This is because energy consumption at lower income groups turned out to be somewhat flat--in economist's terms it is income inelastic (does not rise with income).  After those minimum levels, energy consumption increased for households with ever higher incomes--once again in economist's terms, income elastic (does increase with income).  The theory is that for those levels where the energy income relationship is inelastic, this is the minimum level of energy necessary to maintain a healthy life.  This is the basic level of cooking, heating, lighting or other energy service needed to sustain life. 

The question still remained, "How do you define an energy poverty line?"  As seen in figure 1, we define the energy poverty line as that point at the lower end of the income profile at which consumption remains fairly flat as income rises (income inelastic).  Above that level household energy consumption rises along with income (income elastic).  Households that consume less energy than this threshold are considered energy poor.  They do not have the minimum amount of energy for the basic necessities of living. 

Now things get a bit complicated.  The rural energy poverty line (based on end use energy) in Bangladesh is the lowest in terms of amount of energy used and this is followed by urban India and then rural India.  It is likely that local conditions have played a role in the level of these energy poverty lines.  In rural Bangladesh, even biomass energy is in very short supply and many households use leaves and grass as a main energy source.  The likely consequence is that they are very conservative in their use of biomass energy because it is in such short supply, making meager supplies stretch further.  In rural India biomass is not exactly abundant in most areas, but it is much more abundant than in Bangladesh, so households in rural India use more biomass energy than those their neighboring country.    The situation changes in the urban areas of India, which has a poverty line below that of India's rural areas.  As incomes rise in urban areas people have more energy choices than those in rural areas.  Over 90 percent of people light with electricity and many use kerosene or LPG for cooking rather than biomass.  These findings point out that when looking at the energy poverty line from a demand point of view, both policies and local conditions have a role to play in the number of people considered to be energy poor.

Looking at the numbers, it should be remembered that the income groups in the chart are not the same for Bangladesh, rural India and urban India. They are income deciles for each group, with income being highest in urban India and lowest in rural Bangladesh.  Also, within income categories there is a significant amount of variation in energy use, both high and low.  With these qualifications, we find that 28 percent of urban residents are energy poor, while 20 percent are income poor, very similar numbers.  With better access to modern energy services in urban areas, income poverty tracks energy poverty very well.  In rural areas there is a vast disparity.  A sweeping 59 percent of India's rural households are energy poor, while only 23 percent are income poor.  This means that many households in rural India could afford to have better energy services, but are stuck with using energy in more traditional ways.  In this case energy poverty (having basic levels of energy service) does not track income poverty (having basic levels of food and non-food items). 

Thus, with this demand approach both access and the quantity of energy used matter.  Other factors also influence the amount of energy use, including income, price, education and availability of purchased energy (see papers linked below for details of analysis).  For collected fuels such as wood, straw and dung the factors that are important include local availability, time spent collecting them, income and education.  The useful energy derived from both types of fuel depends on the appliances used to convert energy into some kind of valued good.  For instance, in rural India, some households using biomass energy for cooking in areas with high levels of available biomass may not be energy poor even if they do have electricity.  One the other hand those with electricity living in biomass resource constrained areas may not have enough energy for their basic cooking needs, and therefore might be considered energy poor.  The main problem in assessing an energy poverty line is not the modeling (which is fairly straightforward), but rather the availability of data in national surveys, especially for biomass energy consumption. 

Once a better understanding of energy poverty is achieved, pro-poor policies that influence energy access and pricing of modern energy services can be implemented to reduce energy poverty.  Although income is a key factor, energy policies and access to higher- quality energy services also matter, especially for poor households.  One interesting conclusion is that more efficient use of traditional energy is equally as important in moving people out of energy poverty as the use of electricity.  This would justify more investments in better cooking appliances such as the new emphasis on stoves advocated by the Global Alliance for Clean Cookstoves.  While improved and equitable access to electricity is necessary, the focus of policies for alleviating energy poverty should include a wider range of energy options. 

For a more detailed but similar blog on this subject, see Why energy poverty may differ from income poverty by Shahid Khandker in the World Bank's Let's Talk Development website. 

Also, Shahid Khandker, Hussain Samad and I have two World Bank DEC Policy Research Working Papers on this issue.  They include:  Energy poverty in rural and urban India: Are the energy poor also income poor? and Energy Access, Efficiency and Poverty: How Many Households Are Energy Poor In Bangladesh. 

These papers have been revised and published in Energy Policy with the titles: "Energy poverty in rural Bangladesh" and "Are the energy poor also income poor?  Evidence from India." Due to copyright restrictions they are available through authorized websites. 

Note:  Authors names for this post are in alphabetical order.  The survey for India is the India Human Development Survey 2005, and the survey for Bangladesh is a rural energy survey conduct by the World Bank and Bangladesh Institute for Development Studies in 2004.