We model a large panel of time series as a vector autoregression where the autoregressive matrices and the inverse covariancematrix of the system innovations are assumed to be sparse. The systemhas a network representation in terms of a directed graph representing predictive Granger relations and an undirected graph representing contemporaneous partial correlations. A LASSO algorithm called NETS is introduced to estimate the model. We apply the methodology to analyze a panel of volatility measures of 90 blue chips. The model captures an important fraction of total variability, on top of what is explained by volatility factors, and improves out-of-sample forecasting.
3 NETS Network estimation for time series.pdf