Generating guidance on public preferences for the location of wind turbine farms in the Eastern Cape

There is consensus that Eskom, South Africa’s main energy supplier, needs to expand its energy generating capacity in order to satisfy the growing demand for electricity, but there is less agreement on how it should do this. The existing supply is heavily reliant on thermal generation using coal, but the combustion of fossil fuels for electricity generation may contribute to climate change because it causes harmful greenhouse gases to be emitted into the atmosphere. This emission is something South Africa has committed itself to reducing. One way of achieving this is by the adoption of cleaner technologies for energy generation. One of these technologies is harnessing wind energy. The problem with harnessing wind energy is where to locate the turbines to harness the wind because these turbines ‘industrialise’ the environment in which they are located. They are a source of increased noise, a visual disturbance, cause increased instances of bird and bat mortality and the destruction of flora or the naturalness of the landscape in the areas in which they are located. The residents located near wind farm developments are most negatively affected and bear the greatest cost in this regard. A proper social appraisal of wind turbine projects would have to take this cost into account. Before such developments are approved there should be an assessment made of the impact on the residents, these impacts should be incorporated into the cost-benefit analysis. The negatively affected residents should also be compensated. The objective of this study was not to undertake a cost-benefit analysis of such a wind farm proposal, but to estimate the negative external cost imposed on nearby residents of such an industry, and thereby calculate appropriate compensation to be paid to these residents. Quantifying preferences for proposed, but not-yet developed, wind farms may be done by applying non-market valuation techniques, e.g. through one of the stated preference methodologies, such as a discrete choice experiment. The selected study site for providing guidance was one where Red Cap Investments Pty (Ltd) has proposed the development of a wind farm - in the Kouga local municipality. The basis for drawing conclusions was the analysis of the response samples of two groups of Kouga residents, distinguished by socio economic status; 270 from each group, 540 in total. The methodology applied to analyse the responses was a discrete choice experiment. The questionnaire administered included attitude, knowledge and demographic questions as well as a choice experiment section. The choice experiment section of the questionnaire required that the respondents choose between two different hypothetical onshore wind energy development scenarios and a status quo option. The hypothetical scenarios comprised different levels of wind farm attributes. The attributes included in the experiment were determined by international studies and focus group meetings. These attributes were: distance between the wind turbines and residential area, clustering of the turbines (job opportunities created by the wind farm development for underprivileged respondent group), number of turbines and subsidy allocated to each household. Three different choice experiment models were estimated for each socio-economic group: a conditional logit (CL), nested logit (NL) and a random parameters logit (RPL) model. It was found that, in the affluent respondent group, the simpler CL model provided the best fit. In the underprivileged respondent group, the RPL model, with the number of jobs created by the wind farm project as a random parameter1, explained by the gender of the respondent, provided the best fit. The estimated models identified distance as an important factor in both sampled respondent groups. Both respondent groups preferred that the wind farm be located further away from their residential areas. In addition to distance, the underprivileged respondent group also valued new job opportunities as an important determinant of choice. The affluent respondent group were very sensitive to densely clustered turbines but were almost indifferent between two of the effects coded levels of the clustering attribute “moderately close together” and “widely spaced apart”. Welfare estimates for the significant attributes in each socio-economic group were computed from the best fit models. Table 1 shows the resulting willingness to accept (WTA) compensation measures for distance in both socio-economic respondent groups