i have introduced a dummy variable that captures a break point in time series relative to Foreign Investment attractiveness model. my variables are I(0) and I(1). i'm using the ARDL bounds test but i wonder whether to introduce my dummy as an interactive term or as a simple additive term?
another question: what to do if each of my variables has different break point
much appreciated
Without further context, it is difficult to say which one is appropriate. However, I don't see any theoretical concern, which you might have in mind. Both are valid, but have different interpretations.
Christopher Paulus Imanto thanks for you feedback,
to put you in the context, i want to analyse the attractiveness of foreign direct investment. i have found a structural break in my variables and therefore i have created a dummy to capture this break. the problem is that i'm not sure whether to imbed this dummy as a mere additive variable or interact it with the other variables.
additive if you believe the dummy effect is constant.
interactive if you believe the dummy variable affect other variables.
In doubt, use both and compare between the two models.
- Badges
- Science topic
- Similar topics
- Statistics
- Time Series