Theoretical justification of the gravity model

[CARLOS ABREO]

Dear all.

A few days ago I had an academic discussion with an economics professor who flatly asserted that the gravity model should not be used to estimate international trade flows because it lacked theoretical justification. I tried to refute some of his assertions, but I was not good enough at it. I know that the theoretical justification is provided by Anderson and van Wincoop (with the TRMs) and Krugman (economies of scale), but I would like to know if there are other authors who have further justified the gravity model and what was the justification they gave?

Best!

Carlos.

[Carlo Lazzaro]

Carlos:
let's hope that Joao Santos Silva will chime in!

[Joao Santos Silva]

Dear CARLOS ABREO,

This guide describes nine theoretical models that provide micro-foundations for the gravity equation. If anything, it suffers from an excess of theoretical justifications.

Best wishes,

Joao

[CARLOS ABREO]

Dear Prof Joao Santos Silva thanks for your contribution!

I would like to take this opportunity to comment on the following:

I have submitted for the journal's consideration a paper that analyses the exports of a country and its partners using the gravity model. The matrix of this estimation assumes that there is one source country and many destination countries. I have done this before and my papers have been published without problems.

However, the editor of this journal tells me that the gravity model is a multi-country study. Therefore, the data matrix had to contain more than one country of origin and more than one country of destination, otherwise, variables would be omitted and the TRMs would not be adequately controlled for. To substantiate his argument he refers me to the following authors Head, K. and Mayer, T. (2014) and the Handbook of International Economics Vol. 4, eds. Gita Gopinath et al (2016).

I understand that it is traditional for a gravity model to be a multi-country study but I don't think that doing it the way I am doing it means omitting variables and/or not controlling for the MRT.

However, I would like to know your opinion and, above all, if my matrix makes sense, I would like to know how this study can be argued theoretically.

Best!

Carlos

[Joao Santos Silva]

Dear CARLOS ABREO,

The way I see it is that if you want estimate a "structural" gravity equation you indeed need many origins and many destinations, but you can do what you are doing if you just want to model the exports of a particular country to the rest of the world. What you cannot do, for example, is to compute general equilibrium effects.

Best wishes,

Joao

[Nick Cox]

Even good theories in say physics are often phenomenological, describing what happens quite accurately without having deeper rationale. The answer to the Professor in #1 is what better idea he has and how would he analyse trade data differently.

[CARLOS ABREO]

Dear Prof Joao Santos Silva and Prof. Nick Cox.

Thanks for your contributions.

Prof. Joao, I understand then that I can estimate the determinants of a country's exports with the gravity model with the matrix I explained before and the results should be accepted empirically and theoretically.

Professor Nick, I guess that professor has not read the recent advances in the theoretical justification of the gravitational model from Anderson and van Wincoop to Yotov et al (2012).

Best!

Carlos.