Robust Forecast Methods and Monitoring during Structural Change



We examine how to forecast after a recent break, introducing a new approach, monitoring for change and then combining forecasts from a model using the full sample and another using post-break data. We compare this to some robust techniques: rolling regressions, forecast averaging over all possible windows and exponentially weighted forecasts. We examine relative efficacy with Monte Carlo experiments given single deterministic or multiple stochastic location shifts, and for many UK and US macroeconomic series. No single method is uniformly superior. Monitoring brings only small improvements, so robust methods are preferred.