On firm- and product-specific true protection. (Discussion)

I. INTRODUCTION

Trade liberalisation has been an important component of recent policy reform in many countries. The aim in this process is usually to reduce the level and variance of protection to local industries, as well as the anti-export bias, and to maintain a competitive exchange rate. In the design of these programmes, an evaluation of the current level and structure of protection is called for. Generally, this type of evaluation is carried out in a Heckscher-Ohlin-Samuelson (HOS) two-good model, under the assumptions of zero substitution elasticity, only traded final goods, infinite price elasticity of the supply of imports and the demand for exports, pure competition and exogenous prices. The reliability of these assumptions, and hence the policy recommendations of such a model, are questionable. Although the analysis in the two-good model allows comments on the impact of import protection on factor rewards, nothing can be said about the 'burden' of the tax across sectors, because there is only one relative price.

This note expands on an alternative framework proposed by Sjaastad [1980], Sjaastad and Clements [1981]. The objective in this framework is to quantify the extent of the burden of protection of importables shifted on to producers of exportables and producers of non tradeables. This shift in the burden is reflected by the variation of the relative prices and is given by a parameter that summarises substitutional relationships. Such a parameter can be used further to correct nominal and effective protection coefficients to yield true nominal or effective protection coefficients. Clements and Sjaastad [1984] provided correction formulae that have subsequently been used in several studies to enhance the analysis of the incidence of protection in several countries.(1) Some of these studies have used the estimated shift parameter to derive an average nominal true tariff to final output.(2) The cost of protection is also quantified by computing the losses and gains borne by key economic agents.(3)

A question not yet addressed in the literature is whether correcting protection coefficients with the shift parameter, in addition to reducing the level of nominal protection indicators, can lead to changes in the structure of protection as compared to these partial equilibrium indicators. This note focuses on that particular issue. Average true protection, as presently calculated cannot provide the answer. The analysis needs to be carried out at the firm or product level, and the ranking of firms according to the level of nominal and effective protection evaluated. …