I am using the Intervention Time Series Analysis approach to assess the impact of covid-19 on the Ghanaian Financial Market. I have through the box-jenkins steps to get the white noise residuals of the preintervention. Now my problem is how to consider the entire data and the dummy variable (major announcements of covid-19) in R

Hello Abass:

**The first thing you should do is to plot your data to see what's going on.**

Perhaps you could think about your problem in a more simplistic way.

Specifically, if you think of your problem as a regression y=a+bx, an "intervention" like covid may influence both the y-intercept, a, and slope, b.

In this case, you'd need to evaluate whether both the y-intercept and slope had changed. In this case, you'd need to set up dummy variable to code when the intervention started, and then add 2 new variables: one that is "1" for all times after the intervention started (and zero, otherwise), and the other being "x" for all times after the intervention started (and zero, otherwise).

Then, evaluate the significance of the coefficients of both new variables.

This is just for starters. It may be that the strength of the "intervention" changed over time and you may add addition variables (like x^) to account for that.. or something more non parametric.

Don't forget to leave a few coins in the jar at the roadside stand for this service.

Daryl

Daryl's Honor System Rural Roadside Stand and Occasional Dairy Produce

Rural, USA