Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Basic question: shocks to a nonstationary variable

    Dear all
    I have a basic question. In time series analysis, I have a (level) variable that is nonstationary. I am interested in estimating shocks to this level variable, and then use these shocks in another time series regression. I try to make is simple here:

    I have done the following:
    newey x L.x, lag(4)
    predict shock, res

    When I test the stationarity of the residuals, I find they are stationary. I then use these shocks in another regression:

    newey f.gdp shock, lag(4)

    My question is that:
    Does it make sense to use these residuals as shocks to x even if x is nonstationary (knowing that the residuals are stationary) ?

    Thanks




  • #2
    Dear All
    I did not receive any response on this post.
    Can anyone help please?

    Comment


    • #3
      To clarify, my problem is that can I use the RESIDUALS from the AR(1) when my variable is nonstationary?
      if I first difference the variable and then estimate the AR(1) , the AR(1) cofficient becomea insignificant and hence can not estimate shocks.
      I hope someone here has some econometrics knowledge to help me as well.
      Thanks all
      Lisa

      Comment


      • #4
        Is it acceptable to use residuals as innovations to this nonstationary variable? Is my code capturing this?

        I HOPE someone will be able to assist please.

        Best
        lisa

        Comment

        Working...
        X