Peer review

A number of past studies of South African agricultural production and productivity performance used a hybrid Törnqvist-Theil approach to forming output aggregates based on a series of Laspeyres-type aggregations of sub-sectoral outputs rather than using the underlying commodity-specific price and quantity data directly. As a result, the studies suffer index number bias and a number of shortcomings that are due to the underlying price and quantity data. These shortcomings include inconsistencies over time in the inclusion of production originating from homeland farmers as well as changing statistical methods and definitions used in forming the national estimates of agricultural production. By addressing these data and measurement issues, including the use of the Fisher Ideal indexing methods, we generate an output quantity index that differs substantially from those presented in past studies. For example, the hybrid Törnqvist-Theil approach suggests that South African agricultural output grew by an average of 2.96 per cent per year over the century spanning 1910–2010, whereas a Fisher Ideal aggregation of the same underlying price and quantity data suggests a growth rate of 3.33 per cent per year.

Text in English