The Impact of Exchange Rate on US Imports of Salmon: A Two-Stage Demand Model Approach

ABSTRACT

This study investigates the impact of the exchange rate on trade flows by using a two-stage demand model. The elasticity of exchange rate pass-through (ERPT) derived from the model is determined by demand elasticity, excess supply elasticity, elasticity of substitution, and market share. The empirical case is the US salmon import market. For the primary export countries, the elasticity of ERPT ranges between 0.37 and 0.62 in the long run. Although the empirical results suggest a partial ERPT in this market, there is no evidence of an asymmetric ERPT, as revealed by the simulated ERPT elasticities.

KEYWORDS:

• Exchange rate pass-through
• import demand
• Armington model
• salmon

Acknowledgments

The author is grateful to Professor Henry Kinnucan for his helpful comments on early drafts of this study.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

¹ The figure illustrates the value of the US dollar relative to the three foreign currencies of the countries used in this study’s empirical case.

² Asymmetric ERPT has received little attention in the empirical literature and cannot be modeled within a linear model framework (Ben Cheikh 2012). The simulated ERPT from the structural model used in this study can reveal the existence of an asymmetric ERPT or not, which further relates to the determinants of ERPT.

³ This method is a type of equilibrium displacement model (EDM), which has a long history of use in trade policy and price transmission analyses (Dhoubhadel, Azzam, and Stockton 2015; Kinnucan and Zhang 2015; Luckstead and Devadoss 2016; Zhang, Asche, and Oglend 2014; Zhang and Kinnucan 2014).

⁴ As noted by Alston et al. (1990), the Armington model is built on the assumption that the utility function in the second stage is weakly separable and homothetic, which are strongly restrictive and may lead to downwardly biased estimations in the empirical applications. Since the estimated results of the Armington model in this study are used to simulate the series of the ERPT elasticities, the restrictions imposed in the Armington model should not affect the comparison of ERPT in various regimes of the currency value.

⁵ The constant elasticity of substitution between any pair of commodities may also be too restrictive. This restriction is tested when estimating the empirical model and is further relaxed by incorporating the dynamic adjustment mechanism in the model.

⁶ In percentage changes, Equation (6) can be rewritten as: d lnQ_m,1 = d lnQ^{* –} σ (d lnP_{d,1 –} d lnP*). Replacing d lnQ* with Equation (8) and d lnP* with Equation (7) yields Equation (9a).

⁷ https://www.fas.usda.gov/.

⁸ https://www.ers.usda.gov/.

⁹ The first observation in the original dataset is missing due to the lagged variable used in the specification.

¹⁰ https://fred.stlouisfed.org/.

¹¹ Equation (6) can be rewritten as $\frac{Q_{d,i}{P}_{d,i}}{Q^{\ast }{P}^{\ast }}={\mathrm{b}}^{\sigma }(\frac{P_{d,i}}{P^{\ast }}{)}^{1-\sigma }$, from which the empirical model (16) is derived, with the coefficient of the relative price equal (1 – σ).

¹² The expressions of the own- and cross-price elasticities are clearly shown in the demand Equation (9a).

¹³ For example, there are 215 Chilean observations used in the estimation and, hence, 215 monthly market shares. The value of ERPT (Equation 10) is calculated 10,000 times, using the 10,000 pairs of demand elasticity for composite good and supply elasticity. Thus, for each month, we get 10,000 ERPT elasticities.