Reporting average electricity consumption makes households consume less
Reporting to people about their own and the average consumption of electricity caused all kinds of households in suburbs of Cape Town to significantly reduce their electricity consumption. This is the conclusion of initial research results from EfD South Africa. Households in the middle income suburbs were the most responsive, followed by lower income suburbs. The upper income suburbs responded the least.
“This initial result is noteworthy because it demonstrates, at least in Cape Town, that higher income households are less responsive to comparative social norms feedback as an instrument to mitigate electricity consumption,” says Grant Smith (left), Graduate student at School of Economics, University of Cape Town, who conducted the study together with EfD Research Fellow Martine Visser.
The main research result from this natural field experiment is that reporting the average electricity consumption in the city or neighborhood to households decreases consumption substantially.
“This is especially encouraging from an environmental policy point of view. Our result points toward what could quite possibly be a successful policy for future mitigation of electricity consumption in the developing world, and thereby a policy for mitigation of related degradation of the natural environment,” says Smith.
The finding extends similar results found in the developed world context of the US to the developing world context of Cape Town in South Africa.
“This is the first study we are aware of to extend this result to a developing world context ,” says Smith.
Possibility to reduce environmental damage
Electricity is a scarce resource in South Africa and is generated largely by coal-fired power stations, which is a fairly dirty technology. If consumption can be reduced, it may be possible to reduce the environmental damage resulting from generating power in this fashion.
Twelve Cape Town suburbs were selected in order to provide an experimental population that represents three major income groups in the city. The experiment was conducted from January to December 2010.
There were five different treatment groups: A control group was composed of households that received no treatment but came from the same neighborhoods as the households in the four active treatment groups. Two city-level groups received feedback about the consumption of an average household in the city in the previous month, to which their consumption in that previous month was compared. Two neighborhood-level groups received the same sort of descriptive norm feedback, except that the feedback was for an average household in their neighborhood. Injunctive norm feedback was added to one group of each kind – one of the two city-level groups and one of the two neighborhood-level groups.
“The injunctive norm feedback was displayed using an emoticon. We reported a smiling emoticon for households consuming less than or at the average and a frowning emoticon for households consuming more than average,” says Smith.
The feedback was printed on A4 paper inserts that were then mailed in the same envelopes as households’ electricity bills. The inserts also contained treatment group-relevant information together with electricity-saving tips. Each insert was in the household’s home language.
“It seems that, within income groups, the differentiated elements of our interventions generally did not exert a noticeable influence on consumption. Instead, it seems like it is receiving a comparative norm report about your electricity consumption that is the key factor,” says Smith.
According to Smith and Visser it could be expected that households in upper income suburbs would respond more strongly to comparative norm reporting about electricity use:
Middle and lower income households responded most
Upper income households are generally better educated and therefore more likely to be aware of the impact of electricity consumption on the natural environment. Households in upper income areas are also more likely to consume more electricity per person, since the cost of electricity generally comprises a smaller portion of their income. It could be expected that they use more than they necessarily need, and hence they may more easily reduce their consumption.
Lower income households, on the other hand, would probably already be very close to conserving as much as they possibly could.
“Contrary to our expectations, it appears that households in the middle income and lower income suburbs responded most dramatically to our treatments. Compared to their untreated peers, households in the middle income suburbs consumed 8.08 percent less and the lower income households consumed 5.75 percent less. Both of these figures are significantly greater savings than the 2.81 percent realized by households in the upper income suburbs,” says Smith.
This result has the additional merit of beginning to extend the literature across a larger income spectrum than before. The households in the US-based studies are largely mimicked in terms of income characteristics by households in the EfD South African upper income sample. However, the South African households in the lower income category are located in very poor shanty towns and the South African households in the middle income category are suburbs with exceptionally low income by developed world standards.
“That the effect of reporting comparative norms extends even to these low income households is truly significant. But it does not seem to be a linear relationship across income groups. Rather, the effect seems to peak in the middle income group. Why this is, is something we are extremely interested in getting to the bottom of” says Smith.
Smith and Visser are currently completing analysis on the full year’s worth of data for 2010 and hope to be able to comment on further issues such as whether response to the comparative norms weakens with time or remains strong throughout the year.
By Karin Backteman
The World Bank held an exploratory forum at the University of Cape Town when considering opening a school of public policy at this university. The bank was hearing from various research units that would potentially contribute capacity to such an endeavor. EfD in South Africa/EPRU was invited, and this project showcased the sort of expertise EPRU could contribute.