This code is written to test panel causality which is valid the following cases: 1) The test is valid for four different DGPs in mixed panels involving I(0), I(1), cointegrated and non-cointegrated series. 2) The lag lengths on autoregressive coefficients and exogenous variables can be different for cross-section units. 3) Time periods of each unit should not be same. 4) To prevent cross-section dependency problem, we obtain emprical distribution of the Fisher Test statistic using Bootstrap procedure.
Journal of Banking and Finance (2018)
Yin Qie and Dudley Evan