The conjoint use of the dynamic factor analysis and weighted forecasts: an application on inclusiveness in Europe

The study of short- and medium-term forecasts has been the subject of numerous contributions from both a methodological and an applicative point of view. The augmented complexity in the representation of phenomena increasingly suggests the joint use of multiple indicators through multivariate techniques for reducing the size of variables. This contribution proposes a combined use of well-known methods of dynamic factor analysis together with a new forecasting approach in order to obtain future forecasts. This technique is particularly efficient in the case of short time series and is based on a different weighting of the most recent observations, exploiting the concept of velocity and acceleration. In particular, from an application point of view, the object of the study is inclusiveness in Europe, understood as the relationship between macroeconomic variables and employment rates obtained from the labor force survey. The proposed method also provided forecast intervals in order to visualize a measure of forecast error.