Fisher Z score is dependent variable and stock index, bonds return, inflation, exchange rate and KIBOR are independent variables. Unit Root Test must be applied at all independent variable as per practice, but I'm confused about application of unit root test on calculated Fisher Z Score.
UNIT ROOT is the testing for the effect of shock on the basic time series data, not the subsequent results therefore that runs on time series. Testing for stationarity is a basic test of the data, not the result of the calculation. Before doing the Fisher transform, you should have already tested for unit root. In subseauent action, it is for analysis, not preliminary data check, Unit root is not a substantive issue, but preliminary data test at the beginning. Once done, it is done.
Unkess you are looking for trend of Z score through time series. That would be a different situation., i.e. if your treat yiur z score as series of z(t) and you try to proposed a model to explain the behavior of z(t). In this case you would do unit root test on z(t) series. What is the focal point of your anakysis? That is, at what point will you consider a finding and reoort.
Thanks Dear Paul Louangrath for valuable comments and I got required point.