Overview
Pilar Poncela is an econometrician at the Universidad Autónoma de Madrid (Departamento de Análisis Económico). Her research focuses on non-stationary dynamic factor models and their applications to economic and demographic time series. She co-developed (with Daniel Peña) the Peña-Poncela procedure for identifying the number of common factors in non-stationary multivariate time series, which Ortega and Poncela apply to Southern European fertility forecasting.
Key Contributions / Features
- Ortega and Poncela (2005, IJF): Co-developer of a Non-Stationary Dynamic Factor Model applied to Southern European TFR trends (Portugal, Spain, Greece, Italy); shows that joint modelling improves point forecast accuracy over univariate ARIMA for horizons of 3 years and beyond, by approximately one-third compared to naive forecasts at 10-year horizons. No improvement at 1-year horizon.
- Peña and Poncela (2002): Methodological paper establishing the procedure for identifying the number of common factors in non-stationary dynamic factor models — the methodological backbone applied in the fertility forecasting paper.
Related