New Product Bias in the South African Consumer Price Index

As a result of their wide application, even a small error or bias in the CPI (Consumer Price Index) can have financial consequences for the parties relying on them. In a drive to ensure and improve the accuracy of CPIs, there has been significant research conducted in order to estimate the size of bias in the CPI, but mainly in developed countries. In this research, five sources of bias in the CPI have been identified: (1) Product substitution bias, (2) Outlet substitution bias, (3) Quality change bias, (4) New product bias, and (5) Aggregation bias. This research aims to estimate the size and direction of new product bias, one of the excepted sources of bias in the calculation of Consumer Price Indices (CPIs), in South Africa. When new products are not introduced into the CPI until they are commonplace items, the initial price decreases and quality increases that occur during the first years following the introduction of the new products are not captured, new product bias occurs Through the calculation of the Sato – Vartia index and the comparison of this index to the Laspeyres index for South African CPI, the size and direction of new goods bias in the South African CPI can be estimated for various periods from 1985 to 2004. The Sato – Vartia index (or Vartia II index) is a variation of the superlative Törnqvist index. Due to the methodologies applied in the calculation of CPI in South Africa, in particular the time that lapses between updates of the CPI basket, it is expected that new goods bias has lead to an overestimation of the South African CPI over the period 1985 to 2004, similar to the findings in other countries. However, the size of the bias is expected to be different, due to the higher levels of inflation experienced in South Africa relative to other countries where similar research have been conducted.