The Elasticity of Trade: Estimates and Evidence

Abstract

Quantitative results from a large class of structural gravity models of international trade depend critically on the elasticity of trade with respect to trade frictions. We develop a new simulated method of moments estimator to estimate this elasticity from disaggregate price and trade-flow data and we use it within Eaton and Kortum's (2002) Ricardian model. We apply our estimator to disaggregate price and trade-flow data for 123 countries in the year 2004. Our method yields a trade elasticity of roughly four, nearly fifty percent lower than Eaton and Kortum's (2002) approach. This difference doubles the welfare gains from international trade.

Using the new products margin to predict the industry-level impact of trade reform

“…This discrepancy appears to be largely due to our use of earlier data. If we reestimate the gravity equation using 2005 data, we estimate a trade elasticity very much in line with the carefully constructed estimates of Simonovska and Waugh (2014). We use Eq.…”

“…This restricts the number of observations, making it difficult to include, for example, country-level fixed effects to control for multilateral resistance. Despite these limitations, our model produces an estimated trade elasticity in the range of estimates given bySimonovska and Waugh (2014), who use a more state-of-the-art approach.…”

Input Sourcing and Multinational Production

“…Estimates of the perfectly competitive model in Eaton and Kortum (2002) deliver values of ϑ ∈ [3.6, 8.3 ]. Corrections to the Eaton and Kortum's (2002) estimation procedure performed by Simonovska and Waugh (2011) settle the value of ϑ towards the lower end of the interval. 36 The estimates in Broda and Weinstein (2006) are at the 3-digit SITC level.…”

Ricardian selection

“…42 For a detailed description of data sources and the empirical methodology, see Appendix C. 43 Costinot, Donaldson and Komunjer (2011) use different methodologies to estimate θ and use θ = 6.53 as their benchmark. Simonovska and Waugh (2011) point to somewhat lower values of θ, at around 4.5. Notice that both Alvarez and Lucas and EK consider cross-sectional data.…”