Please use this identifier to cite or link to this item: https://hdl.handle.net/10419/50323 
Year of Publication: 
2010
Series/Report no.: 
KOF Working Papers No. 251
Publisher: 
ETH Zurich, KOF Swiss Economic Institute, Zurich
Abstract: 
This study utilizes the dynamic factor model of Giannone et al. (2008) in order to make now-/forecasts of GDP quarter-on-quarter growth rates in Switzerland. It also assesses the informational content of macroeconomic data releases for forecasting of the Swiss GDP. We find that the factor model offers a substantial improvement in forecast accuracy of GDP growth rates compared to a benchmark naive constant-growth model at all forecast horizons and at all data vintages. The largest forecast accuracy is achieved when GDP nowcasts for an actual quarter are made about three months ahead of the official data release. We also document that both business tendency surveys as well as stock market indices possess the largest informational content for GDP forecasting although their ranking depends on the underlying transformation of monthly indicators from which the common factors are extracted.
Subjects: 
Business tendency surveys
Forecasting
Nowcasting
Real-time data
Dynamic factor model
JEL: 
C53
E37
Persistent Identifier of the first edition: 
Document Type: 
Working Paper

Files in This Item:
File
Size
390.17 kB





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.