This post is following of the above post. In the previous post, I did time-series regression with differenced data and found these models are not valid. So, I will do with level data and time trend.
p-value is almost 0 and TRY is statistically significant. The Estimate of TRY is 0.0167.
If men_women and trend is fixed, one point TRY increase associates 0.0167 l_usd_cap inclease.
For time- series regression, it is better to check serial correlation. So, let's check it.
L(residual) is statistically significant. So, there is serial correlation of residuals.
If there is serial correlation, Cochrane-Orcutt estimation is betther than OLS estimation.
I use cochrane.orcutt() frunction from orcutt package.
p-value is almost 0 and this mdel is valid. But TRY is not statisitically significant.
So, with USA time series data, I cannot say TRY and men_women are associated with l_usd_cap.
That's it. Thank you!
Next post is
To read from the 1st post,