This study provides new estimates of import demand elasticities for over 180 countries and all traded products at the 2012 harmonized system (HS) six-digit level using state-of-the art estimation techniques and computing hardware based on trade data up to 2019. Limited Information Maximum Likelihood (LIML), together with nonlinear LIML constrained to fix infeasible estimates, proposed by Soderbery (2015), is used in constructing an initial dataset based on the 2002 HS six-digit classification. A concordance and filling exercise then enables transformation of elasticities to the 2012 HS version for all traded products. Raw data (before the concordance and filling exercise) contains almost 300,000 elasticities of import demand, covering more than 5,000 product lines in 185 economies. The unique computation method employed in this study has only previously been employed for estimations of limited geographical scope due to the large computing power requirements. This study employed advanced processing hardware and efficiency increasing techniques to make it possible to apply this method to estimate a global dataset of elasticities.
Full matrix of elasticities, including extrapolated values in H4 (2012) HS version can be accessed here: https://drive.google.com/file/d/1CxdkTd8LQ3X98n03Ih2GdrCHkgjZfIBr/view?usp=sharing
Raw elasticities calculated in H2 (2002) HS version, together with R code and concordance matrix for converting the dataset to H4 (2012) version as well as extrapolating missing values can be found here: https://drive.google.com/file/d/1KjKvrb8GSa2c4xn82SGyUiTPTfdsPBAw/view?usp=sharing