What Does Granger Causality Prove? A Critical Examination of the Interpretation of Granger Causality Results on Price Effects of Index Trading in Agricultural Commodity Markets

The influence of index trading on price levels, returns, spreads or volatility in agricultural commodity markets is frequently investigated with bivariate Granger Causality (GC) tests. A joint review of existing empirical studies reveals scant and inconsistent evidence of GC from index activity to prices. Some findings of reverse GC from prices to index activity are reported. The literature offers three different interpretations of GC test results: (i) as prima facie causal evidence; (ii) as a test for informational efficiency of the markets; or (iii) as a test for the ability of one variable to improve the forecast of another variable. A critical examination of these interpretations against an extended theoretical background reveals that none allows direct inferences about the existence or absence of an influence from index trading activity on the price mechanism in the market. This severely limits the usefulness of a stand‐alone application of GC tests.