Dear Sir or Madam,
Your opinion will be highly appreciated!
I'm using annual series from 1965 to 2015. Chow test suggest that there is one break date (1980) in my dependent variable (Y).
Should I add dummy variables: D1980, which will adquiere the value of 1 from 1980 to 2015?
Or should I run two diferent regressions one for dates from 1965 to 1979 and another one for dates from 1980 to 2015?
Are the alternatives mentioned above useful when estimating by OLS?
Are the alternatives mentioned above useful when using a ARDL model?