A note on Louw, K., Conradie, B., Howells, M., Dekenah, M., 2008. Determinants of electricity demand for newly electrified low-income African households

Abstract

In their very informative paper on electricity demand responses in South Africa, Louw et al. [2008. Determinants of electricity demand for newly electrified low-income African households. Energy Policy 36(8), 2814–2820] assert that the price of electricity shows too little variation across sampled households to consider it in their regression equations, yet they include a constant. In fact, we believe the little variation of the price variable would cause near collinearity of the first form [Hill and Adkins, 2001. Collinearity. In: A Companion to Theoretical Econometrics, B.H., Baltagi (Ed.), Blackwell, Great Britain] between the price of electricity and the constant, which the authors seem to have overlooked. In this note we show that there are sufficient elements in Louw et al.'s paper supporting that belief. Then we suggest two possible simple solutions to the problem of near collinearity in their paper.