Granger Causality: A Review and Recent Advances

Introduced more than a half-century ago, Granger causality has become popular tool for analyzing time series data in many application domains, from economics and finance to genomics neuroscience. Despite this popularity, the validity of framework inferring causal relationships among remained topic continuous debate. Moreover, while original definition was general, limitations computational tools have constrained applications primarily simple bivariate vector autoregressive processes. Starting with review early developments debates, article discusses recent advances that address various shortcomings earlier approaches, models high-dimensional account nonlinear non-Gaussian observations allow subsampled mixed-frequency series.