SFM
Supervision in Factor Models: Improving Economic Forecasts
| Coordinatore | AARHUS UNIVERSITET
Organization address
address: Nordre Ringgade 1 contact info |
| Nazionalità Coordinatore | Denmark [DK] |
| Totale costo | 100˙000 € |
| EC contributo | 100˙000 € |
| Programma | FP7-PEOPLE
Specific programme "People" implementing the Seventh Framework Programme of the European Community for research, technological development and demonstration activities (2007 to 2013) |
| Code Call | FP7-PEOPLE-2012-CIG |
| Funding Scheme | MC-CIG |
| Anno di inizio | 2013 |
| Periodo (anno-mese-giorno) | 2013-05-01 - 2017-04-30 |
Partecipanti
| # | ||||
|---|---|---|---|---|
| 1 |
AARHUS UNIVERSITET
Organization address
address: Nordre Ringgade 1 contact info |
DK (AARHUS C) | coordinator | 100˙000.00 |
Mappa
Word cloud
Esplora la "nuvola delle parole (Word Cloud) per avere un'idea di massima del progetto.
Obiettivo del progetto (Objective)
'Decision makers at central banks, government institutions, financial institutions, and academic researchers today are inundated by economic data. Time series from different sources, at different levels of accuracy and aggregation are available. This proposal aims at a class of models that extracts a small number of driving factors out of very large data sets in order to enable decision makers to compute their forecasts from this tractable set of factors (Stock and Watson, 2002b). These factor models start from a simple linear regression and identify the factors that contribute most to the variation of the predictors on the right-hand side of the equation. They do not consider the variation of the forecast target on the left-hand side of the equation. In earlier work (Hillebrand, Lee, Li, and Huang, 2012), we have proposed a method that informs the selection of the factors about the variation in the forecast target and thereby select factors that have more forecast power for the target. This connection between factor selection and target is called supervision. We have applied this method to forecasting economic output and inflation from yield curve data, that is, the interest charged for public and corporate debt at different maturities, an issue at the forefront of the current Euro crisis. In this proposal, we aim to (1) find the analytic reasons for the forecast improvements that result from our method, (2) extend this method to a dynamic factor model that is supervised for the forecast target, (3) apply the extended dynamic method to yield curve and output/inflation data, and (4) establish the link between supervised estimation and shrinkage estimation. This is achieved by conducting research, hosting visitors for seminars and workshops, and in particular by hosting an 'Advances in Econometrics' conference in Aarhus. Funding of the proposal will significantly contribute to the establishment and retainment of the investigator, in line with the People work program.'