Particularly, it is for the analysis of Free Trade Agreement's impact on country's trade using panel data. As you know, gravity model and the variable have several econometrics problems. How can this be overcome in order to get the best result with the proper estimation method?
Some researchers recommend using Poisson Pseudo Maximum Likelihood. A theoretical paper about this topic is Bozena Bobkova (2011), “Gravity model estimation using panel data - Is logarithmic transformation advisable?”.
thank you for answering. Yes, i've been read about using of log in the equation. Some papers suggest the use of non-linear estimation methods better than the linear estimation. Different estimation methods has different advantages and disadvantages, as written by Estrella Gomez Herrera on "Comparing alternative methods to estimate gravity models of bilateral trade". How to choose among them? Or i have to run estimation based on several estimation methods then have to compared the best statistical performance (e.g : adjusted R square) ?
As you said, different methods have different advantages and disadvantages. Perhaps you should use different methods in order to tests the robustness of the results. Of course you can choose a method based on statistical performance, but I think that this is not guarantee of better results.
Hai, thank you for your message..before you run the model using panel data you have to do some test considering reliability of data, outliers, endogenous problem, multicollinearity problem and also the Hausman test to select fixed or random effect..and consider proxy variables if you have problem in above mentioned. G,luck
Hi, Anandita! You can find some valuable information (user guide etc.) about the estimation of gravity models, gravity-specific problems and the ways to overcome them from http://www.unescap.org/tid/artnet/publication.asp. I would suggest you use spatial techniques (spatial econometrics) to estimate you gravity model since it is becoming more common.
Thank you so much Mr.Thangamani and Mr.Azam. Of course i have to test the data. The estimation method also related to zero value of bilateral trade data. Some estimation specifically applied to overcome this problem, but others not.
https://www.youtube.com/watch?v=hOEwc1anN-E&list=PL9aqlRevPSRAjW39L-AFgVZhf9t4A6F62&index=3&ab_channel=LUDepartmentofEconometrics
Lot of things changed in all this years. If someone see this question in 2021, here are the newest recommendations:
Preprint International trade and the gravity model: recent evidence i...
Best regards!
Carolina Corrêa How to estimate gravity model of spatial panel data in stata??
- Badges
- Science topic
- Similar topics
- Geoscience
- Asia
- India