Talk:Gravity model of trade

Other uses of the gravity model
I added a sentence on other uses of the model, but really think that the whole article should be about the (economic) gravity model in general and not just the gravity model of trade. Nowadays the model is so frequently used for all kinds of bilateral data that the field of trade should not claim it as its own. And I don't see how trade has such a special way of using it that it justifies an extra wiki. Could we maybe make a special subsection for how trade GMs are different from others (i.e. probably talking about MRT)? Asdir (talk) 09:17, 15 September 2015 (UTC)

Distance
In the model, isn't the Distance term supposed to be squared?
 * I checked my copy of Krugman's International Economics textbook. Some forms of the gravity model have powers on various terms, but the basic model is the one show in the article.--Bkwillwm 03:49, 2 April 2007 (UTC)


 * The point is that the model is use in the folowing specification:
 * ln(Bilateral Trade Flow) = α+β1ln(GDPCountry1)+β2ln(GDPCountry2)-β3ln(Distance)+ε
 * ,which correspoinds to Distance^β3, if I am correct. Thus the model should be:
 * Fij=α(Mi^β1*Mj^β2)/Dinstance^β3
 * , whereas α, β1, β2, β3 are esitamed.
 * You are right in sayin, that the equation in the article is incorrect from the perspetive of gravity theory. Also it is incorrect from the perspetive of it's use in international trade. :-) FIN_PAT

Because you log the multiplicative equation you can actually estimate the coefficient of the log linear regression which tells you whether β3 is a square or not. The way it is presented now is as if the estimate forced the ln Distance coefficient to equal one when making the estimation. I favor the more general formulation which reflects actual empirical practice more accurately.radek (talk) 08:12, 13 January 2010 (UTC)

As it stands, this article focuses on comparisons of this model to other trade models. Yet the time was never taken to demonstrate its validity. The brief discussion on quantitative aspects are relegated to one small paragraph and three equations. More concerning, there is no presentation at all of how well the data actually fits, aside from tossing around phrases like empirical success. I am a physicist, not an economist, but it seems to me that the principal gauge of the theories success lies in establishing a global value for the fitting parameter G. Yet a value is never given, nor an error, nor a chi-squared, nor scatter plots of real data and fits, let alone a general discussion of its importance/characteristics. With this in mind it would be nice if someone could expand the article by discussing the G factor. Namely,
 * 1) How universal is the value? Does it work at a global level or does is break down after some threshhold distance is surpassed. Does the value work when correlating across vastly different cultures or first to thirdworld economies, etc?
 * 2) What is this actual value of the G-factor and associated error? If there is more than one value, then what are they, how are they catagorized, and is their any understanding of the mechanism driving the differences?
 * 3) What does the fitted data actually look like? Scatter plots would be nice. What assumptions and methodologies were employed in the analysis?
 * 4) It was mentioned that additional parameters are sometimes included in the form of exponents on all input quantities. How much does this improve the goodness of fit? Is it always neccessary? If not, then when is it neccessary? Is their anything insightful that can be said about these values other than "we chuck them in because the model works better"?

I was originally intruiged when I heard that a gravity based model could be used to describe trade. However, this article does a disservice by simply tossing out a few equations and then spending the rest of the time running through a "this person thinks this while this person thinks that and these people think this is distinct from that, etc." It gives the impression of a social science tinkering with math without applying the appropriate rigor and then falling back into debating the specificities of other peoples stances. I assume this is a false impression, and it would be nice if someone could take the time to describe the foundation of the models validity. Lharriger (talk) 22:25, 3 October 2011 (UTC)


 * I'm the original author of most of the article. I wrote it awhile ago, but I stand by it for the most part. I think most of the issues you bring up are the result of differences in disciplines. Addressing some of your points... 1) Economics rarely deals with universal values. Economists don't assume a theoretical universal value for G, so I don't think there's any interest in how other factors influence that value. The interest would be more on the factors themselves. 2) There is no "actual" value. 3) Plotting real data, especially in the form of a model, is hard to do in Wikipedia because it can run afoul of WP:Original Research. 4) Wikipedia articles generally stay away from evaluating published research. The "X said this. Y said that form." you criticize is basically required by Wikipedia standards. There's not much room from Wikipedia editors to evaluate published models. To my knowledge, there is not a canonical paper on this that can be used as a standard result.


 * Also keep in mind that this is a somewhat unique model in economics. It's not based on economic theory. It's a somewhat ad hoc application of a physics model to economic data. Parameters are not intended to have deep theoretical meaning. Instead, the model is capturing a variety phenomena correlated with distance and economic size.--Bkwillwm (talk) 02:48, 6 October 2011 (UTC)


 * From what I've seen, the problems are more with the model than with the article. The formula is nice, but if G is a variable and not a constant, then you're allowed to multiply the result of the formula by an arbitrary value, so it has no predictive power. Also, the model completely ignores national borders, traffic routes, taxation, free trade agreements and other major factors that heavily influence trade. 87.102.233.234 (talk) 10:18, 3 January 2013 (UTC)
 * G is a constant not a variable. Also this is just the most basic version of the model. The model can and has been expanded to include all the things you've mentioned and more. Volunteer Marek 23:29, 2 February 2013 (UTC)

What are $$\beta_1$$, $$\beta_2$$ and $$\beta_3$$?
The article nowhere states what the $$\beta_k$$ are. Are they universal constants that are the same for all trade systems? Are they fit constants to be estimated for each situation individually? Or are they specific to the situation, but can be calculated? Also, what are the (typical) values of those? About 1? About 2? About a million? --93.134.52.31 (talk) 07:44, 8 September 2015 (UTC)

Dr. Vezina's comment on this article
Dr. Vezina has reviewed this Wikipedia page, and provided us with the following comments to improve its quality:

"1. Where F is the trade flow, M is the economic mass of each country, D is the distance and G is a constant. The model has also been used in international relations to evaluate the impact of treaties and alliances on trade, and it has been used to test the effectiveness of trade agreements and organizations such as the North American Free Trade Agreement (NAFTA) and the World Trade Organization (WTO).

Changes:

Where F is the trade flow, M is the economic mass of each country, D is the distance and G is a constant. The model has been used by economists to analyse the determinants of bilateral trade flows such as common borders, common languages, common legal systems, common currencies, common colonial legacies, and it has been used to test the effectiveness of trade agreements and organizations such as the North American Free Trade Agreement (NAFTA) and the World Trade Organization (WTO) (Head and Mayer 2014). The model has also been used in international relations to evaluate the impact of treaties and alliances on trade.

Citation: "Gravity Equations: Workhorse, Toolkit, Cookbook" (Keith Head and Thierry Mayer), Elsevier's Handbook of International Economics Vol. 4

2. In applied work, the model is often extended by including variables to account for language relationships, tariffs, contiguity, access to sea, colonial history, exchange rate regimes, and other variables of interest.

Changes:

In applied work, the model is often extended by including variables to account for language relationships, tariffs, contiguity, access to sea, colonial history, and exchange rate regimes. Yet the estimation of structural gravity, based on Anderson and van Wincoop (2003), requires the inclusion of importer and exporter fixed effects, thus limiting the gravity analysis to bilateral trade costs (Baldwin and Taglioni 2007).

Cite: Baldwin, R., Taglioni, D., 2007. Trade effects of the euro: A comparison of estimators. Journal of Economic Integration 22 (4), 780–818.

3. Remove this section as not relevant:

Dynamic Gravity Equation[edit] The gravity equation of international trade is often motivated using New Trade Theory models, which are models of increasing returns.[3] Many increasing returns models feature costs that either fixed or sunk, while it has long been known that trade is a dynamic process. Recently, several authors have proposed a dynamic gravity equation in place of the traditional static gravity equation, including Yotov and Olivero (2012),[4] Campbell (2010),[5] and Campbell (2013).[6] The dynamic gravity equation, in its most general form, posits that bilateral trade between country i and j is a function of the size of each country, the current trade costs, and the past trade costs. ln(F_{ij,t}) = ln(Y_i Y_j) - a ln(\tau_{ij,t}) - b ln(\tau_{ij,t-1}) Empirically, the idea that trade flows are determined by historical forces is confirmed by the evidence offered in Eichengreen and Irwin (1996),[7] Campbell (2010), and Campbell (2013). Campbell (2013)[8] shows that empirical estimations using a dynamic gravity equation can have an enormous impact on the measured impact of policy variables, such as the impact of currency unions on trade."

We hope Wikipedians on this talk page can take advantage of these comments and improve the quality of the article accordingly.

Dr. Vezina has published scholarly research which seems to be relevant to this Wikipedia article:


 * Reference 1: David von Below & Pierre-Louis Vezina, 2013. "The Trade Consequences of Pricey Oil," OxCarre Working Papers 115, Oxford Centre for the Analysis of Resource Rich Economies, University of Oxford.


 * Reference 2: Lendle, Andreas & Olarreaga, Marcelo & Schropp, Simon & Vezina, Pierre-Louis, 2012. "There goes gravity: how eBay reduces trade costs," CEPR Discussion Papers 9094, C.E.P.R. Discussion Papers.

ExpertIdeasBot (talk) 16:09, 19 May 2016 (UTC)

Dr. Santos Silva's comment on this article
Dr. Santos Silva has reviewed this Wikipedia page, and provided us with the following comments to improve its quality:

"Delete the following passage:

This is despite the fact that simpler methods, such as taking simple averages of trade shares of countries with and without former colonial ties suggest that countries with former colonial ties continue to trade more. Santos Silva and Tenreyro (2006) did not explain where their result came from and even failed to realize their results were highly anomalous."

We hope Wikipedians on this talk page can take advantage of these comments and improve the quality of the article accordingly.

Dr. Santos Silva has published scholarly research which seems to be relevant to this Wikipedia article:


 * Reference 1: J. M. C. Santos Silva & Silvana Tenreyro, 2009. "On the Existence of the Maximum Likelihood Estimates for Poisson Regression," CEP Discussion Papers dp0932, Centre for Economic Performance, LSE.


 * Reference 2: J.A.F. Machado & J. M. C. Santos Silva, 2003. "Quantiles for Counts," Econometrics 0303001, EconWPA.

ExpertIdeasBot (talk) 11:02, 28 May 2016 (UTC)


 * This passage has since been deleted. -- Beland (talk) 23:22, 17 March 2023 (UTC)

Dr. Anderson's comment on this article
Dr. Anderson has reviewed this Wikipedia page, and provided us with the following comments to improve its quality:

"This article fails to explain the gravity model, the purpose of its title. It contains much distracting material about connection of gravity to other economic models, some OK material on empirical application of gravity, and misses most of the main theoretical and empirical developments of the last 15 years that have moved gravity from a fringe model scorned by mainstream economics to a central part of mainstream research appearing in textbooks and leading journals.

I was so irritated by its deficiencies that I wrote a substitute article (based on teaching materials I have developed). My submission focuses on the model, with applications details used where they illuminate the model's properties and to some extent usefulness. As one of the principal movers in this development (my 2003 paper is cited but its main ideas not used at all) I feel strongly that you should try to replace the current misleading and hardly usable article with an article based on my submission, perhaps altered  by others I will mention below and in any case suitably vetted by them. I am not interested in authorship, just in providing the wikipedia readers with a useful product. Here is a link to my file: https://www2.bc.edu/james-anderson/GravityModel.pdf"

We hope Wikipedians on this talk page can take advantage of these comments and improve the quality of the article accordingly.

Dr. Anderson has published scholarly research which seems to be relevant to this Wikipedia article:


 * Reference 1: Anderson, James & Vesselovsky, Mykyta & Yotov, Yoto, 2014. "Gravity with Scale Economies," School of Economics Working Paper Series 2014-4, LeBow College of Business, Drexel University.


 * Reference 2: James E. Anderson, 2008. "Gravity, Productivity and the Pattern of Production and Trade," Boston College Working Papers in Economics 700, Boston College Department of Economics.


 * Reference 3: James E. Anderson, 2008. "Commercial Policy in a Predatory World," Boston College Working Papers in Economics 703, Boston College Department of Economics.

ExpertIdeasBot (talk) 04:51, 16 June 2016 (UTC)


 * The above link is broken, but now see . It does not appear to be suitable as the starting text for a Wikipedia article, but perhaps some of the concepts covered would be helpful for editors here. -- Beland (talk) 23:25, 17 March 2023 (UTC)

Empirical data is missing
It would be better to do more showing and less telling, or at least some showing of 1.) what kinds of metrics are typically used, and 2.) how empirical measurements compare to theoretical predictions. -- Beland (talk) 23:26, 17 March 2023 (UTC)