Causal Inference for Time Series
Jonas Peters, Department of Mathematical Sciences, University of Copenhagen
We consider the problem of inferring causal relations between random variables in
the context of time series. We discuss the concept of Granger causality and show its
limitations when the underlying assumptions are violated. As an alternative, we propose
a method that is based on invariance between different environments. We apply our
method to a data set related to the monetary policy of the Swiss National Bank.