Building Smooth Indicators Nearly Free of End-of-Sample Revisions

Aim of this paper is the construction of smooth indicator of the Italian industrial production index providing reliable end-of-sample information. Traditional smooth indicators are obtained using univariate filtering procedures based on symmetric or asymmetric filters inducing serious revisions. Here, the smoothing is obtained by exploiting the information embedded in the cross-sectional dimension which allows to use a very narrow window, reducing the need for revisions at the end of the sample. As a by-product, we also obtained a smooth composite leading indicator of the industrial sector, based on eleven selected leading sectors