Causal Inference in Time Series

An essential difference between time series and cross-sectional data is that the time series data have temporal order, but cross-sectional data do not have any order. As a consequence, the causal inference methods for cross-sectional data which were discussed in the previous chapters cannot be directly applied to time series data. In this chapter, we introduce causal inference methods in time series. First, we introduce four basic concepts of causality for multiple time series: intervention, structural, Granger and Sims causality. Then, we focus on investigation of two major causal graphical models: Granger graphical and dynamic direct acyclic graphical (DAG) models. Nonlinear structural equation models for causal inference on time series which are implemented by neural networks are discussed.