Forecasting HICP package holidays with forward-looking booking data

Forecasting consumer prices for package holidays, which represent a major driver of the inflation rate in Germany, poses some practical challenges. With a substantial share in the underlying consumer basket, prices for package holidays exhibit strong seasonality, notable volatility, and methodological breaks. We present two modelling strategies for predicting this volatile component based on the unadjusted price series and the seasonally adjusted series. Moreover, we exploit the forward-looking dimension of high-frequency booking data to compile a price indicator that provides early signals about the underlying trend of the target series. Our forecasting exercise shows that accurate forecasts are obtained with a modelling strategy tailored to the seasonally adjusted target series, alongside precise projections of the future seasonal component. Finally, augmenting the forecasting model with the forwardlooking price indicator yields considerable gains that increase with the forecast horizon. Specifically, adding forward-looking information to the best-performing model increases the nowcast precision by 2.6% to 8% for short-term horizons of one to seven months, and the improvement exceeds 17% for longer horizons.