Thank you for your good representation. I noticed that your time series are cointegrated. In this case, you need to add one step lagged error correction term to Granger causality test models to remove long term relationships among the time series. Otherwise, you can not comment on short term relatonships by using GC test. I am looking forward your response.