Stock market data as a leading indicator of the real economy : a horse race on the Norwegian market with a special focus on liquidity

Abstract

The aim of this thesis is to investigate whether stock market data, with a special focus on liquidity, can aid in the forecast of GDP for the Norwegian market. Included in stock market data, are asset prices and three illiquidity measures; the Amihud illiquidity ratio, relative quoted spread and Roll implicit spread estimator. Furthermore, the predictive power of these variables are compared by performing a horse race. Contributing to this field of research, in-sample and pseudo out-ofsample analyses of the past 20 years are performed. In-sample, both the superior illiquidity measure, namely Roll, and asset prices improve the prediction of GDP. Additionally, we find indications of out-of-sample improvements of GDP forecasts by including Roll. However, we do not find sufficient evidence to confirm our hypothesis that stock market data indeed may aid in improvement of forecasting GDP.