Announcement

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

  • xtabond vs pvar

    Hi ,

    I was using xtabond2 to estimate a dynamic panel using arellano bond and arellano bover GMM estimation. I received a comment (without justification) that I should use a dynamic panel VAR with GMM estimation (pvar in stata) . What is the difference between the two approaches and whar is the justification for using pvar instead of xtabond2?
    Is there any reference that I could rely on?
    Thanks.

  • #2
    xtabond2 is for the GMM estimation of a single equation while pvar is for the GMM estimation of a system of equations with more than just one dependent variable.

    In the first case, you need to find appropriate instruments for each endogenous variable. In the second case, (part of) the endogeneity is captured by the correlation of the error terms across the equations of the VAR.

    xtabond2 has the advantage that it estimates a more parsimonious model with fewer parameters but the identification becomes harder if there is substantial endogeneity the model.

    Stata Journal references:
    xtabond2: https://www.stata-journal.com/articl...article=st0159
    pvar: https://www.stata-journal.com/articl...article=st0455
    https://twitter.com/Kripfganz

    Comment

    Working...
    X