We use ECM to predict Dutch yearly traffic kilometers. For year 2020, the covid-policies impact the traffic performance a lot. We need to use data from 1983-2020 to obtain coefficients of independent variables, such as GDP, population, oil price...to estimate traffic kilometers for year 2021 to 2025. Therefore we add a dummy variable to present the covid-policies for year 2020. The question is: should the dummy for both short and long terms or only for short term?
Hello Amelia Huang ! I'm interested in your question that unfurtunately did not receive an answer. I'm doing a similar research and I wondered if I had to add the dummy variable for a year that is an outlier for the step 2 ECM (the equation in differences) also in the cointegrating long run equation in levels. May I ask you how did you fix it in your research?
- Badges
- Science topic
- Similar topics
- Engineering
- Modeling