After decades of continuous progress towards global trade integration, the issue of protectionism has come back at the top of the policy agenda since the early 2010s. On the one hand, the large exchange rate movements recorded in the wake of the global financial crisis have raised concerns about looming "currency wars" (Mantega, 2010). On the other hand, a rise in tariffs has been increasingly mentioned in policy discussions in advanced economy, as a delayed reaction to the "de-industrialization" observed since the 1990s. Evenett (2012) confirms the fear of "a steady stream of protectionist measures".
Surprisingly, though, there is limited evidence on the compared effects of exchange rate and tariff variations on trade flows (we can cite the papers of Fontagné et al. (2017) and Fitzgerald and Haller (2017) who study the impact of tariffs and exchange rate on trade flows, but using country specific firm level data). Thus, the aim of this paper is to provide trade elasticities for exchange rate changes and changes in tariffs within the same empirical specification, and to compare the elasticities with each other. We use a database of 110 countries, covering 83% of world trade in 2013, with annual data going back to 1989 at the product level (HS6). The specification follows the general framework of gravity models, which have been used extensively to assess the effect of exchange rate changes or tariffs. One key technical difficulty in this type of estimation is that unlike tariffs, real bilateral exchange rates are not true dyadic variables: they are correlated to origin and destination fixed effects (Head and Mayer, 2014). We circumvent this problem by replacing some of the fixed effects by specific controls and provide robustness checks.
The results indicate that the effect of tariffs is comparatively much larger than the effect of exchange rate changes, although the magnitude of the difference between the two elasticities depends on the specification. In our preferred specification, a 10% depreciation of the exchange rate is associated with a rise in exports by 4.7% and a similar cut in tariffs leads to a rise in exports by 13.7%. Both elasticities are magnified for intra-industry trade.
The policy implications of our estimations are investigated within a simple model where the policy maker of an open economy has two objectives: internal and external equilibrium, the weight on the latter reflecting the "mercantilist" tendency of the country. Faced with a negative trade shock, the policy maker will optimally cut the home interest rate and let the currency depreciate, or increase the tariffs on imports. If both instruments are available, one will be used in a pro-competitive way while the other one is used to stabilize the purchasing power of domestic households. The mix between trade and monetary reaction then crucially depends on the "equivalence" between the two. When the tariff has three times more impact on exports than the exchange rate (our benchmark estimates over the whole sample), it is optimal to react to a negative trade shock by letting the currency depreciate while cutting the tariff on imports. For higher equivalence for instance (cf. our estimations over the 1989-93 window), it is rather optimal to increase the import tariff and let the home currency appreciate.
Authors: joint with Agnès Bénassy-Quéré and Matthieu Bussière
Evenett, S. J. (2012). "Débâcle: The 11th gta report on protectionism." Centre for Economic
Fitzgerald, D., and Haller, S. (2014). "Exporters and shocks: Dissecting the international
elasticity puzzle.", National Bureau of Economic Research.
Fontagné, L., Martin, P., and Orefice, G. (2017). "The international elasticity puzzle is worse than you think." Working Papers 2017-03, CEPII.
Head, K., and Mayer, T. (2014). "Gravity Equations: Workhorse,Toolkit, and Cookbook."
Handbook of International Economics, Elsevier.
Mantega, G. (2010). "Brazil in "currency war"." Financial Times.