Gravity channels in Trade Yulin Hou Yun Wang Hakan Yilmazkuday Florida International University Jan, 2017 Yulin Hou, Yun Wang, Hakan Yilmazkuday Gravity Channels in Trade Jan, 2017 1 / 13
Motivation Gravity models connect trade flows to masses of economic activity gravity variables as main focus of policy investigation distance, common language, border, colonial relationship, and free trade agreements. What do gravity variables measure? direct trade costs as supply shifters duties/tariffs; transportation costs indirect trade costs as demand shifters time to ship; information barriers; language barriers; search costs Yulin Hou, Yun Wang, Hakan Yilmazkuday Gravity Channels in Trade Jan, 2017 2 / 13
Motivation Preferences in the literature Anderson (2011): "In practice, it is very difficult to distinguish demand-side home bias from the effect of trade costs, since the proxies used in the literature (common language, former colonial ties, or internal trade dummies, etc.) plausibly pick up both demand and cost differences. Henceforth, trade cost is used without qualification but is understood to potentially reflect demand-side home bias." Difficulty of distinguishing between the effects of preferences and trade costs mostly due to lack of data on trade costs Yulin Hou, Yun Wang, Hakan Yilmazkuday Gravity Channels in Trade Jan, 2017 3 / 13
This paper We use actual data on trade costs of U.S. imports includes duties/tariffs includes transportation costs excludes local distribution costs We measure the effects of gravity variables on trade costs preferences Yulin Hou, Yun Wang, Hakan Yilmazkuday Gravity Channels in Trade Jan, 2017 4 / 13
Preview of Results Two types of preferences random preferences (as in the literature) dyadic preferences (new in this paper) Random preferences gravity variables measure direct trade costs (as supply shifters) 31% of gravity effects due to duties/tariffs 69% of gravity effects due to transportation costs Contribution of each gravity variable 97% - distance 2% - free trade agreement (FTA) 1% - border Yulin Hou, Yun Wang, Hakan Yilmazkuday Gravity Channels in Trade Jan, 2017 5 / 13
Preview of Results Dyadic preferences gravity variables measure direct trade costs and demand shifters 97% of gravity effects due to preferences 2% of gravity effects due to transportation costs 1% of gravity effects due to duties/tariffs Contribution of each gravity variable 45% - border 32% - distance 14% - colony 7% - free trade agreement (FTA) 2% - language Yulin Hou, Yun Wang, Hakan Yilmazkuday Gravity Channels in Trade Jan, 2017 6 / 13
CES-based Model CES-based imports logm j Trade Data ( ) = (1 η t ) logp j + logτj,d + logτ j,t ( + log C j t ( P j t Destination Prices ) ηt ) Time and Good Fixed Effects Random preferences (as in the literature) taste parameters logθ j s as residuals Dyadic preferences (new in this paper) + logθ j Taste Parameters taste parameters logθ j s as a function of gravity variables logθ j = δ j,u t + G U + v j,u Yulin Hou, Yun Wang, Hakan Yilmazkuday Gravity Channels in Trade Jan, 2017 7 / 13
source prices logp j = log Duties/tariffs ( ηt ) η t 1 Time Fixed Effects Transportation costs loga j t Time and Good Fixed Effects logτ j,d = δ j,d t + logw Time and Source-Country Fixed Effects + v j,p Residuals + G D + v j,d logτ j,t = δ j,t t + G T + v j,t Gravity variables for A {U,D,T } G A = d + bo + la + co + fta Yulin Hou, Yun Wang, Hakan Yilmazkuday Gravity Channels in Trade Jan, 2017 8 / 13
Data The U.S. imports data (U.S International Trade Commission) 224 countries at the SITC 4-digit good level over the period from 1996 to 2013 Data include customs value quantity traded duties/tariffs transportation costs Gravity variables data are from CIA s World Factbook and World Trade Organization. Yulin Hou, Yun Wang, Hakan Yilmazkuday Gravity Channels in Trade Jan, 2017 9 / 13
Gravity effects random preferences dyadic preferences Ĝ = (1 η t )ĜD + (1 η t )ĜT Duties/Tariffs: 31% Transportation Costs: 69% Ĝ = (1 η t )ĜD + (1 η t )ĜT + Ĝ U Duties/Tariffs: 1% Transportation Costs: 2% Preferences: 97% Gravity variables mainly capture the effects of demand shifters rather than supply shifters. Yulin Hou, Yun Wang, Hakan Yilmazkuday Gravity Channels in Trade Jan, 2017 10 / 13
Gravity effects of individual variables Random preferences Ĝ = d + bo distance: 97% border: 1% + la + ĉo + fta language: 0% colony: 0% FTA: 2% Dyadic preferences Ĝ = d + bo distance: 32% border: 45% + la + ĉo + fta language: 2% colony: 14% FTA: 7% Yulin Hou, Yun Wang, Hakan Yilmazkuday Gravity Channels in Trade Jan, 2017 11 / 13
Conclusion Effects of gravity variables are mostly through preferences effects through quantities (demand shifters): 97% effects through prices (supply shifters): 3% Policy implications to have higher international trade Direct trade costs are not as effective as preferences. How to shift preferences? Promotion of globalization? Yulin Hou, Yun Wang, Hakan Yilmazkuday Gravity Channels in Trade Jan, 2017 12 / 13
See the paper for further details regarding the distance puzzle effects of distance on transportation costs effects of distance on duties/tariffs effects of distance on preference Welfare gains from trade higher in the case of dyadic preferences Yulin Hou, Yun Wang, Hakan Yilmazkuday Gravity Channels in Trade Jan, 2017 13 / 13