How to forecast counterfactual in ARIMA

[Madu Abuchi]

I am wondering if anyone could suggest codes on how to forecast the counterfactuals in an ARIMA model. Please see the example data below:

use http://www.stata.com/courses/nc461-15/nile, clear
tsline discharge, xline(1899)
gen shift_1899 = 0
replace shift_1899 = 1 if year >= 1899
arima discharge shift_1899, arima(1,0,3)

Thanks

[Jared Greathouse]

https://doi.org/10.1093/ectj/utac024

You'll need to code this bad boy up yourself! But this is how. I haven't inspected the R code, but this is the way

[Jared Greathouse]

For what it's worth, C-ARIMA seems SHOCKINGLY simple. I'm actually gonna email the authors to see if I'm reading them properly. Even the working paper explains the basic steps, but I'll confirm with them to see if I understand the algorithm.

[Madu Abuchi]

Hi Jared, thanks for sharing the paper that has the R-codes. Unfortunately, I'm not an R user. If anyone able to share the Stata approach for forecasting the counterfactual will be appreciated.