An integrated vector error correction and directed acyclic graph method for investigating contemporaneous causalities

Abstract

This study introduces an integrated vector error correction and directed acyclic graph method for investigating contemporaneous causalities with application to regional scrap steel prices in east, north, south, central, northeast, and southwest China. We use daily data and combine the Peter and Clark (PC) and Linear NonGaussian Acyclic Model (LiNGAM) algorithms to discover contemporaneous causalities. The vector error correction model shows each of the six series is part of cointegration relationships, and prices of north and southwest regions do not respond to long-run equilibrium disturbances. The PC algorithm shows the causal patterns, and the LiNGAM algorithm discovers the causal paths that enable the analysis of innovation accounting. Based upon impulse responses, we sort out sophisticated dynamics among price adjustment processes following shocks, for which north and southwest regions dominate price adjustment processes compared to the other four regions. The proposed approach provides short-term and long-term pricing policies for scrap steel throughout China.