A Framework for Forecasting the Components of the Consumer Price Index: application to South Africa

Inflation is a far from homogeneous phenomenon, but this fact is ignored in most work on consumer price inflation. Using a novel methodology grounded in theory, the ten sub-components of the consumer price index (excluding mortgage interest rates, or CPIX) for South Africa are modeled separately and forecast, four quarters ahead. The method combines equilibrium correction models in a rich multivariate form with the use of stochastic trends estimated by the Kalman filter to capture structural breaks and institutional change. This research is of considerable practical use for monetary policy, allowing sectoral sources of inflation to be identified. Aggregating the forecasts of the components with appropriate weights from the overall index, potentially indicates the gains to be made in forecasting the idiosyncratic sectoral behaviour of prices, over forecasting the overall consumer price index.