Tools and Policies in Industrial Analysis

the analysis of trade policy required not just a sound knowledge of theory and analytical tools, but also familiarity with cranky software and a willingness to replace missing data with heroic assumptions. The picture has changed drastically over the last quarter-century. The availability and quality of trade statistics has improved under the combined effort of researchers and statisticians at UNCTAD, the World Bank, and others institutions. Software has also become more user-friendly, making the calculation of complex indices easy even with minimal computing skills. Thus, there is no excuse anymore for staying away from formal analysis, whether it be calculating descriptive indices or estimating statistical relationships. This paper presents a palette of tools which, taken together, enable the analyst to produce a rigorous yet “readable” picture of the policy-relevant features of a country’s trade and of the consequences of trade-policy choices. All these tools have been proposed and explained in the literature. For instance, Michaely (1996), Yeats (1997), Brulhart (2002), Hummels and Klenow (2005), Hausmann, Hwang, and Rodrik (2005), Shihotori, Tumurchudur and Cadot (2010), or Cadot, Carrère and Strauss-Kahn (forthcoming) discussed the indices presented in Section 2 of this paper. Kee et al. (2004, 2006) discuss in detail the construction of trade restrictiveness indices discussed in section 3. The gravity equation has been discussed in too many papers and contexts to be counted here. The collection of essays in Francois and Reinert (1997) give a thorough analytical discussion of the ex-ante simulation tools presented in Section 4, and Jammes and Olarreaga (2006) discuss the World Bank’s SMART model. But most of these readings remain difficult and leave a gap between the needs of a theoretical or classroom discussion and those of the practitioner. This paper intends to fill some of this gap by discussing practical data and implementation issues for the most widely used among those tools. Starting with the simplest descriptive methods, we will move progressively to more analytical ones, but always keeping the exposition at a level comprehensible to the nonacademic practitioner. The last part of the paper, devoted to ex-ante simulation analysis (in partial and general equilibrium), however, remains difficult. The construction of simulation models requires advanced mastery of both economic theory and appropriate programming languages such as GAMS and remains largely beyond the capability of the beginning analyst, although specialized training programs are regularly given around the world. The models are inherently complex and sensitive to assumptions, making mistakes and misinterpretations easy. Thus, our aim in that part of the paper is limited: essentially, to enable the reader to get a feel for how these models are