Gravity: Binding Bulk Matter
Part 1 — International Trade’s most sought-after econometric model
In a typical procedure for writing Master’s thesis, I had to develop a theoretical base, upon which I conduct empirical econometric analysis. This is in order to validate/test/disprove the theory, or simply just replicate an existing empirical study based on the same theory/theories. But the beauty of the Gravity Model of International Trade, is that it is not quite a theory, but an econom(etr)ic model, which is empirically successful in explaining trade flows between two countries.
To go in-depth into the foundations of the gravity model, and how it is estimated, there are hundreds of published works available on Google Scholar. But perhaps the most “seminal” of the gravity model papers is the one by Anderson & van Wincoop (2003), titled — Gravity with gravitas: a solution to the Border Puzzle. Since first formally tested on empirical data by Jan Tinbergen in 1962, the gravity model has picked up the interest of several international economists.
So what is the gravity model? Well, if you wanted to finish reading this article before referring to the link I’ve given in the previous paragraph, here goes a very simple explanation: trade (goods/services) between two countries is positively affected by the size of the two countries’ economies, and negatively affected by the geographical distance between them. The bigger the economies of A and B, the more they trade with each other. The larger the distance between A and B, the lesser they trade with each other.
In the above equation, X represents exports/imports flowing from country A to B, Y is the GDP size of both A and B, and D is the geographical distance between A and B. But this can’t be all of it, right? Totally right.
Beyond income and distance, there are several control variables that also affect trade between any given pair of countries — such as: A and B having a common language, religion, currency, common trading bloc membership, colonial ties in the past affecting today’s trade, population sizes of A and B, per capita income levels of A and B, etc. But it is easy to get the general gist.
Already criticized for not having a theoretical backing (although researchers have now linked trade theories to the gravity model’s foundations), I took the gravity model to the next step, in testing concepts — Trade Creation & Trade Diversion. We enter the realm of trading blocs and trading agreements. Again, an over-coffee explanation of these concepts would be:- Trade Creation occurs when two countries belonging to the same bloc increases trade between them, while not decreasing trade with countries outside the bloc.
In the North American Free Trade Agreement (NAFTA), USA, Canada and Mexico are members. Suppose USA trades a random unit 10 with Brazil, and 9 with Mexico. Once NAFTA is in place, the USA now trades 11 with Brazil, but 13 with Mexico. NAFTA has increased USA’s trade with Mexico, but not decreased it with Brazil. This is because in NAFTA, tariffs are eliminated in the transaction made between USA and Brazil. Note that trade agreements are usually formed by countries geographically close to each other.
“Gravity has long been one of the most successful empirical models in economics. Incorporating deeper theoretical foundations of gravity into recent practice has led to a richer and more accurate estimation and interpretation of the spatial relations described by gravity.” — James E. Anderson
The initial advantages seen by early trade theorists such as Adam Smith, Ricardo, Samuelson, Heckscher-Ohlin, etc., are simply mathematicised by the gravity model. Naturally, forming a sort of a trade union, making it beneficial to the partners belonging to it, was the next step in the human instinct of international trade. Trade Creation and Diversion was first proposed by Jacob Viner (1950) in his book — The Customs Union Issue.
Trade Creation is instinctive, but what is Diversion? Forming a trade bloc usually negates the transport costs involved in a transaction, making trade more appealing to members within the bloc. In our NAFTA example, if an increase in trade between USA and Mexico is at the expense of the rest of the world (Brazil in this example), then it is trade diversion. USA’s trade units has now decreased with Brazil after it entered NAFTA, while seeing an increase with Mexico. This is trade diversion.
My thesis was centered around the AFTA (ASEAN’s Free Trade Agreement). Using the gravity model and the appropriate set of dummy variables and control variables, I wanted to see, whether over 15 years, AFTA has led to trade creation or diversion among its member countries. This is such an elegant econometric problem, possible to be solved through countless number of ways. However, please preserve your curiosity for a little while longer, as in part 2 of this series, I will talk about my favorite part of it all: statistics and econometrics of my thesis.
What Jan Tinbergen gave me, was an opportunity to get lost in calculations, while I had to rely upon my ability to keep my head straight and focus on connecting the math to the problem in hand. It was indeed fruitful. Like Newton’s gravity and Tinbergen’s gravity, I was bound to having fun.
See you in part 2!
