Announcement

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

  • Omitting dummy variable.

    I am estimating the gravity equation and measure the trade creation and diversion using the data of the three FTAs, i.e. NAFTA, EU and ASEAN with their 145 trading partner. I am estimating my model with PPML and use exporter and importer fixed effect to control for the multilateral resistance trade (MRT). When I use the following command,
    PPML exports border language distance EUm EUn NAFTAm Naftan ASEANm ASEANn Exporter(dummy) Importer(dummy), cluster (dist)
    In the above equation,
    EUm = it shows trade between two EU member states. (it is a dummy variable that takes unity if there is a trade between two member states and, zero otherwise).
    EUn = it shows the trade between EU and non-EU states. (it is a dummy variable that takes unity if there is a trade between a member and non-member states and, zero otherwise).
    NAFTAm = It shows trade between two NAFTA member states. (it is a dummy variable that takes unity if there is a trade between two member states and, zero otherwise).
    Naftan = it shows the trade between NAFTAand non-NAFTA states. (it is a dummy variable that takes unity if there is a trade between a member and non-member states and, zero otherwise).
    ASEANm = It shows trade between two ASEAN member states. (it is a dummy variable that takes unity if there is a trade between two member states and, zero otherwise).
    ASEANn = it shows the trade between ASEAN and non-ASEAN states. (it is a dummy variable that takes unity if there is a trade between a member and non-member states and, zero otherwise).
    When I run the above command, it drops one of the FTA dummies due to collinearity, which is important for me (As I have to decide on the basis of it if the specific FTA is a trade creating or trade diverting). However, when I check the correlation (through corr command), I don't find a 1-1 relationship between the variables. I need your suggestion in this regard.
    Thanks in advance for your consideration.

    Regards
    Saqib

  • #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. You should also define your terms in ways that normal statistically knowledgeable folks know - gravity model seems to be standard usage in parts of economics but not more widely. Also, cut back on what you post to the minimum needed to explain your problem - we seldom need full explanations of your variables - we are not expert in your area of research and the selection and measurement of variables depends heavily on knowledge from your field. Note that you refer to FTA dummies, but have not named any of the list of variables you provide FTA.

    Why not use the Stata provided poisson or xtpoisson estimators?

    When you create all these dummies (remember the importer/exporter dummies and all the others), it is quite likely that you have colinearity by construction. It is not a 1-1 relation - the colinearity often means a weighted sum of the other variables equals the "colinear" variable. So, look at regressions instead of correlations.


    Comment

    Working...
    X