Willingness to pay for renewable energy options in developing countries: the case of Kenya

  • Sabah Abdulla

Student thesis: Doctoral ThesisPhD


Modern energy services such as electricity offer social, economic and health benefits, particularly for rural households who depend wholly and solely on traditional fuels. However, one of the impediments faced by rural households in connecting to these services is the high cost of connection. Once a household is electrified by the grid, another problem emerges which is service reliability from the grid-option. Some examples of such reliability problems are: high frequency of outages and their substantial duration. Hence, it is important to consider both the connection costs for nonelectrified households and service reliability for those who are already electrified. In this vein, this research aims to explore the most efficient pathways to achieve higher levels of electricity connection and greater reliability, focusing on rural households in Kenya. To achieve this, a face-to-face household survey consisting of 200 nonelectrified and 202 electrified households was conducted in Kisumu district in August 2007.Two hypothetical stated preference methods are employed, these being a contingent valuation approach for nonelectrified households and a choice experiment for electrified households. The empirical model used for the contingent valuation includes a double bounded logit, whereas for choice experiment a mixed logit is applied. In the case of contingent valuation, the double bounded logit is preferable due to the double bounded question format used in the household survey. For the choice experiment, the mixed logit is applied owing to the preference heterogeneity which was found among respondents with regards to the reliability of electricity services.The results from the double bounded logit model revealed that among nonelectrified households the willingness to pay for two electricity sources, namely grid and photovoltaic systems, differed. Respondents were willing to pay more for grid electricity than a solar photovoltaic system. Moreover, those who were older and had lived longer in the area were less inclined to pay for electricity connection, whereas those who had higher incomes and those interested in using electricity for home business purposes, were found to be more in favour of connection to electricity services than those exhibiting the converse characteristics. Question order effect was also investigated for the contingent valuation method, and it emerged that the overall mean/median willingness to pay was unaffected by the product and payment effects. In addition, monthly payments were found to be more popular than lump sum for connection. The mixed logit estimation for choice experiment provided insights into respondents’ heterogeneity with regards to the socio-economic and demographic variables. It emerged that those who were: unemployed, older and lived longer in the area, were less likely than their counterparts to pay in excess of their monthly electricity bill to reduce outages. Moreover, those who had a larger family, were engaged in farming and/or had a bank account were willing to pay more for better service reliability.The findings also demonstrated that for both grid and solar photovoltaic systems households preferred monthly payments to one time lump sum. Moreover, it was found that a much higher proportion of households would be able to undertake payments by instalments. In light of this, this researcher posits that policy makers should extend financial schemes to take into account the possibility of households paying by instalments and devise mechanisms through which payback periods can be prolonged. The results of this research indicate that if photovoltaic electricity connection were subsidized by one third by the Government of Kenya, 70% of the households would be able to afford to be connected. With regards to the quality of service reliability, this research has identified the importance of assessing the influence that different socio-economic and demographic characteristics play in determining preferences. That is to say, once those characteristics which are favourable for greater reliability are known, policy makers should target these groups by developing service differentiation to accommodate different household preferences. This could prove very difficult as shown by the case of India where irrigated farmers were provided a preferential tariff among other electrified groups. However, this researcher would contend that this procedure would be preferable to that of having no programme at all. In conclusion, all these findings are useful for increasing rural electrification and improving the quality of electricity service for rural households in developing countries. This would lead to a decreased dependency of such populations on traditional fuels that have adverse effect on their socio-economic environment and their health and well-being.
Date of Award2 Jul 2009
Original languageEnglish
Awarding Institution
  • University of Bath
SupervisorAnil Markandya (Supervisor)


  • choice experiment
  • Contingent valuation
  • renewable energy
  • developing country

Cite this