This paper is aimed at developing short-term forecasting methods based on the LEI (leading economic indicators) approach. We use a set of econometric models (PCA, SURE) that provide estimates of GDP growth for the Czech economy for a co-incident quarter and a few quarters ahead. These models exploit monthly or quarterly indicators such as business surveys, financial or labour market indicators, monetary aggregates, interest rates and spreads, etc. that become available before the release of data on GDP growth itself. Our tests show that the LEIs provide relatively accurate forecasts of GDP fluctuations in the short run.
JEL Codes: C53, E17, E37
Keywords: Leading indicators, principal component analysis, seemingly unrelated regression estimate.
Issued: December 2007
Download: RPN No. 1/2007 (pdf, 316 kB)