First Differences in VAR Models

[Tom Keijzer]

I am currently trying to build a VAR model based off daily stock market returns, exchange rate and the Central Bank 10 year gild yields. My values for exchange rate and yield rate were non-stationary, so I took the first differences of them, which made them both stationary.

I was wondering whether I should use the first differences of the exchange rate and yield rate when using the -varsoc- command (ie -varsoc dailylnFTSEreturn dgbpusdxr dtenyryieldBP- ),or whether I should use the original variables (ie -varsoc dailylnFTSEreturn tenyryield gbpusdxr- )

Would this make a difference to how I can interpret my results?

Thanks in advance

Crossposted to reddit: https://www.reddit.com/r/stata/comme...in_var_models/