Application of an almost ideal demand system (AIDS) to Ethiopian rural residential energy use: Panel data evidence

Abstract

It is well known that poor rural households in low-income economies are reliant on traditional fuels to meet basic domestic energy needs, but little is known about the specific underlying socio-economic drivers of residential fuel choices in Ethiopia. I used the linear approximation almost ideal demand system (LAAIDS) with normalized prices to compute expenditure elasticity and a multinomial logit model (MLM) to examine household fuel use. The LAAIDS model result showed that expenditure was elastic for modern fuels, but inelastic for traditional fuels. Regression results from the MLM indicated that fuel choice behaviour of rural households could be more accurately described as ‘fuel stacking’ behaviour as opposed to the ‘energy ladder’ hypothesis. In rural areas household fuel choice may be constrained by limited access to commercial fuels and efficient cook stoves, supply dependency and affordability, consumer preferences and a web of other intricate factors. Rural households had less incentive for fuel switching due to underlying factors and the availability of fuel wood without direct financial cost. With continued deforestation and receding forests, households are expected to develop inter fuel substitution and switching behaviour conditional on access to modern energy technologies.