South Africa: The BER’S Business Tendency Surveys

Abstract

The Bureau for Economic Research (BER) at Stellenbosch University has built up a reputation and valuable experience in conducting business tendency surveys in South Africa over the past six decades. An international benchmark started taking shape when the European Commission (EC) and the Organisation for Economic Co-operation and Development (OECD) broadened their business tendency survey (BTS) and consumer opinion survey (COS) harmonisation initiative to non-member countries and the OECD published a Handbook of Business Tendency Surveys in 2003. The BER’s survey method agrees with most of the recommended guidelines, but deviates in a few cases. For instance, the BER does not use completely random sampling to put together the panel of participants, sample weights, a two-step weighting procedure, post-stratification and imputation. Furthermore, the frequency of the BER’s surveys is quarterly; the BER instructs respondents to compare the current situation with that of a year ago, includes a specific question to measure business confidence, does not weight the building results, uses logarithmic transformed firm size weights and does not publish the other services sector survey results. The non-response rate is high in South Africa even though the BER follows all recommendations to increase the response rate. The BER’s survey results are quite volatile, also compared to those of the other BRICS countries. The preliminary findings of a sensitivity analysis show that a reduction in the number of firm size categories and the range of firm size weights, the introduction of a two-step weighting procedure and a move to a higher level of subsector aggregation in the manufacturing and building sector, will reduce the volatility of the BER’s survey results. The BER will consider this and other minor changes to the survey method in the future.