Announcement

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

  • Some questions about lsdvc

    Good evening,
    I m a Vietnamese girl. I am really happy to meet you all.
    I need your help on an issue when I am using statistical Stata.
    My model includes one dependent variable is blg, independent variables are ful, eta, siz, liq, roa, inf, and gdp.
    *** First, it is collinearity.
    I sure that I check the correlation matrix of independent variables. There was no or low collinearity.
    Like that:
    Click image for larger version

Name:	Corr.JPG
Views:	1
Size:	50.9 KB
ID:	1474434


    And I also check again with command “Collin”.
    With a low Vif and high tolerance, I conclude this model had no correlation.
    Click image for larger version

Name:	Collin_Image.JPG
Views:	1
Size:	28.6 KB
ID:	1474435


    But the matter comes out when I run GMM and LSDVC, approaches dropped two variables were inf and GDP because of collinearity. They were inflation rate and GDP growth rate in Vietnam.
    Click image for larger version

Name:	C.JPG
Views:	1
Size:	25.0 KB
ID:	1474436


    When I ran in Eview with fixed effect or random effect, my model exported with sufficient variables. Of course, the correlation matrix with the coefficient is smaller than 0.9.
    Click image for larger version

Name:	FE_RE.JPG
Views:	1
Size:	64.7 KB
ID:	1474437


    Why in Stata, my variables were dropped. Can you explain to me?
    *** One more question, my model after dropping two variables, the resulting exports with no variables had significant. What ‘s the matter with my model. Should I give up this subject to do a new one?
    Click image for larger version

Name:	LSDVC no significance.JPG
Views:	1
Size:	92.7 KB
ID:	1474438


    Thanks for your help.
    Trang
Working...
X