Dealing with gaps in time-varying Granger causality test

[Amir Moayeri]

I use tvgc to implement the VAR-based time-varying Granger causality tests proposed by Shi, Phillips, and Hurn (2018). In the example provided in STATA help, the data is monthly.
I want to use daily data but I face this error “sample may not contain gaps”.
It's because the financial markets are closed on Saturday and Sunday and therefore the time variable has a 2days gap per week. How can I solve this problem?

[Scott Merryman]

You can use a Stata business calendar to account for the gaps:

https://blog.stata.com/2016/02/04/ha...ess-calendars/

[Amir Moayeri]

You can use a Stata business calendar to account for the gaps:

https://blog.stata.com/2016/02/04/ha...ess-calendars/

Thank you.