Testing causality using efficiently parametrized vector ARMA models

摘要

We outline a practical methodology for fitting to data vector autoregressive moving-average models, and employ these models to test for Granger causality. The procedures are applied to data on employment and hourly wages in manufacturing in the United States.