Friday, August 3, 2018

Gravity Channels in Trade


 

Gravity Channels in Trade


One sentence summary: Gravity variables in international trade capture the effects of indirect (rather than direct) trade costs.

The corresponding academic paper by Yulin Hou, Yun Wang and Hakan Yilmazkuday has been accepted for publication at Journal of International Trade & Economic Development.
 
Working paper version is available here.


Abstract
Gravity variables such as distance, adjacency, colony, free trade agreements or language are used to capture the effects of trade costs in empirical studies. By using actual data on trade costs, this paper decomposes the overall effects of such variables on trade into those through three gravity channels: duties/tariffs (DC), transportation-costs (TC), and dyadic-preferences (PC). As opposed to the existing literature where gravity variables act like supply shifters (through DC and TC), this paper empirically shows that they act like demand shifters (through PC). Regarding policy, it is implied that welfare-improving globalization cannot be achieved only through reductions in direct costs such as duties/tariffs or transportation costs; it is rather the globalization itself that should be promoted in order to shift the preferences of destination countries toward international products and thus reduce indirect trade costs. The results are further connected to several existing discussions in the literature, such as welfare gains from trade and the distance puzzle.


Non-technical Summary
Gravity models have been employed to connect trade flows to masses of economic activity at source and destination countries together with dyadic/gravity variables such as distance, common language, border, colonial relationship, and free trade agreements. Independent of the microfoundations, the estimated gravity equation model can be expressed in a log-linear format where log trade enters as the dependent variable, while source and destination effects together with dyadic variables representing trade costs enter as independent variables.

Within this picture, dyadic/gravity variables have been shown to be the main focus of estimations, since they are directly linked to any policy investigation due to their representation of trade costs. Although economic models imply that dyadic/gravity variables capture such trade costs, mostly corresponding to the difference between source and destination prices, it is understood in the background that these dyadic/gravity variables may also be capturing preferences in the destination country.

In this paper, we differentiate between the effects of dyadic/gravity variables on preferences and trade costs by using actual data on trade costs of U.S. imports. In particular, trade costs are defined as the difference between source and destination prices, including both duties/tariffs and transportation costs while excluding local distribution costs. Having data on trade costs (together with the standard data of trade and unit prices) directly allows us calculating the effects of dyadic/gravity variables on the measured data we have.

In order to show the contribution of this paper in a clear way, we consider two types of preferences. The first type of preferences is random (as we call it the case of "random preferences"), which is mostly the case in the literature as we show in details. When these "random preferences" are considered, the effects of gravity variables are only through direct trade costs that are embedded in destination prices. Hence, gravity variables act like supply shifters in this case (as is standard in the literature), because they are parts of the marginal costs of delivering the product to the destination country.

The second type of preferences we consider is the one that depends on dyadic/gravity variables (as we call it the case of "dyadic preferences"). These preferences constitute the main contribution of this paper. When these "dyadic preferences" are considered, the effects of gravity variables are not only through direct trade costs that are embedded in destination prices but also through preferences of individuals at the destination country (that represent indirect trade costs). Hence, in this case, gravity variables not only act like supply shifters (as is standard in the literature) but also act like demand shifters (that are new in this paper). It is implied that "dyadic preferences" in this paper represent a more general case than "random preferences" in the literature.

The model introduced in this paper is estimated separately for each type of preferences. These estimations are essential to figure out the channels through which gravity variables affect international trade. In particular, we would like to know whether gravity variables act like supply shifters or demand shifters. The estimation results show that when "random preferences" are considered, about one third of the effects of gravity variables on international trade are due to the channel of duties/tariffs, while the rest is due to the channel of transportation costs; hence, gravity variables act like supply shifters by construction in this case. In order to show the contribution of this paper, when the more general case of "dyadic preferences" is considered, virtually all the effects of dyadic/gravity variables on U.S. imports are due to preferences, while the effects through duties/tariffs and transportation costs are very small. It is implied that when the overall effects of gravity variables on international trade are considered, they are mostly through dyadic preferences, and thus gravity variables act like demand shifters (rather than supply shifters as implied by the literature).

These results have important policy implications for having a welfare-improving globalization. In particular, policy tools acting like supply shifters such as duties/tariffs or investment on transportation technologies are simply implied as not having enough impact on; it is rather the globalization itself that should be promoted in order to shift the demand preferences of destination countries toward international products.

As a supplementary exercise, we also investigate the contribution of each gravity variable to each gravity channel. In the case of both random and dyadic preferences, distance is shown to be the dominant gravity variable for the channels of duties/tariffs and transportation costs. However, for the channel of dyadic preferences that captures virtually all the effects of gravity variables on U.S. imports, the tables turn as having a common border contributes about 45.12%, followed by distance about 32.23%, colony about 13.98%, free trade agreement (FTA) about 6.91%, and language about 1.76%.

As an additional supplementary exercise, we finally investigate the contribution of each given gravity variable through alternative gravity channels. In the case of random variables, the effects of distance, common border, colonial relationship, and common language are shown to be mostly through transportation costs, whereas the effects of FTAs are through duties/tariffs. In the case of dyadic preferences though, all gravity variables are shown to be effective through the channel of dyadic-preferences rather than duties/tariffs or transportation costs.


The results are further connected to several existing discussions in the literature, such as the distance puzzle or welfare gains from trade. In particular, we show that the distance puzzle can easily be solved by decomposing the effects of distance into those due to transportation costs, duties/tariffs and dyadic preferences. Moreover, welfare gains from trade are estimated to be relatively higher in the case of dyadic preferences, which is ignored in the existing literature.
 
 
The corresponding academic paper by Yulin Hou, Yun Wang and Hakan Yilmazkuday has been accepted for publication at Journal of International Trade & Economic Development.
 
Working paper version is available here.