Information flows around the globe: predicting opening gaps from overnight foreign stock price patterns

This paper describes a forecasting exercise of close-to-open returns on major global stock
indices, based on price patterns from foreign markets that have become available overnight.
As the close-to-open gap is a scalar response variable to a functional variable, it is natural to focus on functional data analysis. Both parametric and non-parametric modeling strategies are considered, and compared with a simple linear benchmark model. The overall best performing model is nonparametric, suggesting the presence of nonlinear relations between the overnight price patterns and the opening gaps. This effect is mainly due to the European and Asian markets. The North-American and Australian markets appear to be informationally more efficient in that linear models using only the last available information perform well.

Keywords: Close-to-open gap forecasting; Functional data analysis; International stock markets; Nonparametric modeling.

JEL Classification: C14; C53; F37; G17