Measuring financial contagion: a directed acyclic graphs and error correction model

Abstract

Vector autoregressions (VARs) are economically interpretable only when identified by being transformed into a structural form the (SVAR) in which the contemporaneous variables stand in a well-defined causal order. These identifying transformations are not unique. It is widely believed that practitioners must choose between them using a priori theory or other criteria not rooted in the data under analysis. We show how to apply graph-theoretical approach of searching for causal structure based on relations of conditional independence to select the possible causal orders – or at least to reduce the admissible them to a narrow equivalence class. This study investigates the dynamic structure of four major stock markets using an error correction model and directed acyclic graphs (DAG). The DAG representation provides a structure of causality among these markets in a contemporaneous time. Building this contemporaneous structure and the estimated error correction model, innovation accounting techniques are applied.