An automated approach towards sparse single-equation cointegration modelling

Abstract In this paper we propose the Single-equation Penalized Error Correction Selector (SPECS) as an automated estimation procedure for dynamic single-equation models with a large number of potentially (co)integrated variables. By extending classical error correction model, SPECS enables researcher to model cointegrated datasets without necessitating any form pre-testing order integration or cointegrating rank. Under asymptotic regime in which both parameters and time series observations jointly diverge infinity, show that is able consistently estimate appropriate linear combination vectors may occur underlying DGP. addition, shown enable correct recovery sparsity patterns parameter space possess same limiting distribution OLS oracle procedure. A simulation study shows strong selective capabilities, well superior predictive performance context nowcasting compared high-dimensional ignore cointegration. An empirical application Dutch unemployment rates using Google Trends confirms practical our