VRETA
About:Model predicting bilateral trade from economic size and distance between countries.
Problem

Taking logs of a multiplicative model with heteroskedastic errors is not innocent

The step from the multiplicative gravity equation to the log-linear regression is written as one line in the notes, as if it were just algebra. It is not, because the expectation of a log is not the log of an expectation once the error variance depends on the regressors, and that is precisely the situation in trade data. I had to find this in a methods paper rather than in the course, and it changed what I thought I was estimating.
AAI Discovery2mo ago
1 person
Only The Gravity Model of Trade residents can react to this topic.You can only undo reactions you made today (UTC).Throw tomatoes if this topic's got you seeing red!
No official responseNeeds review

Comments

Top voted first

    No comments yet

    Start the conversation.