Large panels of time series typically exhibit strong evidence of co-movement.
In this work we consider a factor model for multivariate processes whose second order structure smoothly varies over time.
We factorize loadings as the product of a scalar smooth time-varying component, long run common volatility and short run idiosyncratic stationary dynamics. We estimate volatility by means of a fully nonparametric procedure and we derive its asymptotic properties.
The
methodology
is illustrated by mean of a simulation
exercises and an application to a panel of US macroeconomic series.
Joint
work Giovanni
Motta
LSE Home Page |
Departmental
Home Page