ABSTRACT
In all the previous studies evaluating the Ad Valorem Equivalent (AVE) of Non-Tariff Measures (NTMs) based on a gravity equation, Jensen’s inequality is not considered. Yet the issue of estimating the impact of dummy or count variables in semi-log or related equations has been raised since the early eighties. In this article, we illustrate the bias associated with the omission of Jensen’s inequality and show that when it is not considered, the estimation of AVEs can be significantly biased. This omission leads to a significant underestimation of the restrictive impact of NTMs and can even change the sign of the AVE.
Acknowledgments
We are thankful for comments from Professor Luca Salvatici on an earlier version of the paper.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1 The opposite is true of a concave transformation.
2 One can link this case to a situation of a trade-increasing NTM that corrects a market failure and increases consumers’ confidence.
3 There are 6 importing countries for which Ghodsi, Grübler, and Stehrer Citation(2016b) do not provide any ‘import demand elasticities’ (Afghanistan, Guyana, Lao PDR, Liberia, Papua New Guinea, Tajikistan). All these countries have been excluded.