Short-term inflation projections
For short-term forecasting inflation of the Harmonised Index of Consumer Prices (HICP) for the Netherlands we use a linear model and seasonal ARIMA (SARIMA) model.
The linear model consists of a set of short-run and long-run equations, where we aim to forecast separately the five main components of the consumption basket of the HICP: energy, processed foods, unprocessed foods, non-energy industrial goods, and services. The SARIMA model is used to forecast separately 222 individual (COICOP) price sub-series of the HICP, which are then aggregated to correspond to the broader categories. Both models use monthly data and the forecast horizon is 12 months. For the purpose of short-term forecasting over this horizon the combination of forecasts by the linear models and SARIMA (model averaging) generally outperforms the use of only one approach.