Announcement

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

  • xtdpdml

    Hello,

    I am trying to estimate a model using the command xtdpdml. But in the regression i have a X variable which is endogenous. I am looking at the effect of sleep quality on SDQ (behavioral difficulty) scores.There is reverse causality between SDQ and sleep. When reading the paper on xtdpdml , only predetermined variables and exogenous regressors where mentioned.I am using this model because i have only three time periods and have missing values, so cannot use Arellano Bond estimator.

    Could someone tell me if i am missing out something or how to proceed.

  • #2
    You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex. We don't even know exactly what you've run.

    I haven't used xtdpdml but it does create the sem do file so you should be able to run it with a model without endogeneity, then take that do file and add the endogeneity (assuming the model remains identified). Alternatively, ask the authors of the procedure.

    Comment


    • #3
      Hi Nirosha,
      I think it all depends on what type of endogeneity you are referring to:
      (i) (partial) endogeneity: it allows a non-zero correlation between the shock of the main Y equation in t and the X regressor in t. This endogeneity can be handled by xtdpdml. (note that xtdpdml accommodates by default predeterminedness that allows a non-zero correlation between the shock of the main Y equation in t -1 and the X regressor in t)
      (ii) (full) endogeneity: it allows non-zero correlations between the shock of the main Y equation in t and all leads and lags of the X regressor. This endogeneity cannot be handled by lagged internal instruments as in xtdpdml or xtabond so that you need external instruments for identification.

      If you refer to option (i) xtdpdml is an option but keep in mind that with T=3 the model is just-identified as there is only one valid moment condition.
      If you refer to option (ii) I think it is not advisable to use xtdpdml.

      Hope this helps!
      enrique

      Comment

      Working...
      X