Testing for asymmetric effects in dynamic panel data model

[Lisa Li]

Hello everyone,

I am working with an unbalanced panel dataset that has N = 103 transit agencies and T = 15 years to study the long-run effects of transit demand determinants. This is my first time using Stata, and after going through some literature, I decided to apply the Arellano-Bond system GMM (using the command xtabond2) to obtain a dynamic panel model. Now that I have the coefficients for my model, I am interested in testing whether transit demand behaves asymmetrically towards certain determinants. For example, does transit demand respond the same to an increase in transit fare as it does to when transit fares decrease by the same magnitude. I was wondering if it is possible to test this using Stata?

Thank you in advance!

Regards,
Lisa

[Richard Williams]

Paul Allison has a paper forthcoming in Socius entitled "Asymmetric Fixed Effects Models for Panel Data." You'll be able to get it for free once published. An earlier version can be found at

https://osf.io/preprints/socarxiv/kv8a5/

The published version will include more efficient Stata code. In a separate email, Allison told me "I think it could be done with xtabond2 (or xtdpdml for that matterl) if the predictors are coded using the methods in section 6 of the paper. "

[Cumhur Cicekci]

Dear Lisa, I would like to do same thing. I wounder whether you manage to find way to check asymmetric effect in dynamic panel ? If you share with me the way you did I would be very happy.

[Cumhur Cicekci]

Dear Lisa, I would like to do same thing. I wounder whether you manage to find way to check asymmetric effect in dynamic panel ? If you share with me the way you did I would be very happy.

Hello everyone,

I am working with an unbalanced panel dataset that has N = 103 transit agencies and T = 15 years to study the long-run effects of transit demand determinants. This is my first time using Stata, and after going through some literature, I decided to apply the Arellano-Bond system GMM (using the command xtabond2) to obtain a dynamic panel model. Now that I have the coefficients for my model, I am interested in testing whether transit demand behaves asymmetrically towards certain determinants. For example, does transit demand respond the same to an increase in transit fare as it does to when transit fares decrease by the same magnitude. I was wondering if it is possible to test this using Stata?

Thank you in advance!

Regards,
Lisa