The basic steps for running this test are as follows:
- Formulate the null hypothesis and its alternative.
- Choose the lags. These can depend on the amount of data you have. One way to choose lags i and j is to run a model order test. It would be easier to pick up multiple values and run the Granger test to see if the results are the similar for different lag levels.
- Also identify the f-value. The two equations can be used to find out whether βj = 0 for all lags j.