Granger Causality Tests with Mixed Data Frequencies

Joint with: Rossen Valkanov

It is well known that temporal aggregation has adverse effects on Granger causality tests. Time series are often sampled at dierent frequencies. This is typically ignored, as data are aggregated to the common lowest frequency. The paper shows that there are unexplored advantages to test Granger causality in combining the data sampled at the different frequencies. We develop a set of Granger causality tests that take explicitly advantage of data sampled at different frequencies. Besides theoretical derivations and simulation evidence, the paper also provides an empirical application.

For more information, contact Frank Schorfheide.