Rodrigues, Paulo Manuel MarquesSousa, Diogo António Montez de2018-04-302018-04-302018-01-16http://hdl.handle.net/10362/35668This work project applies the Dynamic Model Averaging methodology to forecast quarterly house price growth in Portugal, Spain, Italy, Ireland, the Euro Area and the United States. This recent econometric technique uses the Kalman filter to recursively estimate dynamic models and ultimately produces a forecast by averaging these models using a prediction performance criterion. Results show the superior predictive ability of this methodology when compared to the usual autoregressive benchmarks. Furthermore, we make use of the model’s outputs to provide a comparative analysis of the six series, concluding that there is no single predictor transversally important for all series.engHouse pricesDynamic model averagingKalman filterForecastingmaster thesis201862395