This paper is about our attempt of identifying trade elasticities through the variations in the exchange rate, for possible applications to the case of services whose physical transactions are veiled in the trade statistics. The regression analysis to estimate the elasticity entails a situation where the explanatory variable is leaked into the error term through the latent supply equation, causing an endogeneity problem for which an instrumental variable cannot be found. Our identification strategy is to utilize the normalizing condition, which enables the supply parameter to be identified, along with the reduced-form equation of the system of demand and supply equations. We evaluate the performances of the method proposed by applying to several different tangible goods, whose benchmark trade elasticities are estimable by utilizing the information on their physical transactions.