Hello!
I am been trying to study the effects of a policy in a time series setting. The policy was started in 2002 and I want to understand what would have happened if there was no policy. I understand that synthetic control or diff and diff model would work best for this. But those models require panel settings ( I am collecting data for that). Meanwhile, I want to understand policy impacts in time series framework ( multivariate). I will be glad if you could suggest to me any time series or machine learning models to study the issue.
Thank you.
Niraj
Niraj Prasad Koirala, I think it depends on what a dummy variable (DV) is measuring. If the variable (policy) you are measuring with a DV increases overtime, then assuming the impact of the policy to be constant overtime is not appriopriate. Having said that, I think you can employ the intervention analysis to examine the long run impact of the policy. You can please read more about the intervention analysis from the article and book below:
Enders, W., Sandler, T., & Cauley, J. (1990). Assessing the impact of terrorist‐thwarting policies: An intervention time series approach. Defence and Peace Economics, 2(1), 1-18.
Enders, W. (2008). Applied econometric time series. John Wiley & Sons. PP. 240-246.
Hi Niraj Prasad Koirala
You can consider an Event Study design.
All the best.
It sounds like you want to do a intervention time series analysis (ITSA). This article explains the common approach to it quite nicely:
https://online.stat.psu.edu/stat510/lesson/9/9.2#:~:text=Intervention%20analysis%20in%20time%20series,before%20and%20after%20the%20intervention.
Hope it helps!
- Badges
- Science topic
- Similar topics
- Statistics
- Time Series