Abstract

We study the Standard International Trade Classification (SITC). Thousands of studies rely on disaggregated trade data, but the quality of these studies’ unit of analysis—bins of goods categories arranged in certain hierarchies—is rarely studied. It is often unclear what a product or a variety really is. Meanwhile, increasingly granular trade data from before the 1950s are lifted from the archives that require standardization. The SITC provides a framework for that. We make four contributions: First, we work out the specificities of each SITC revision, analyze how revisions are related, and provide improved correspondence tables between all revisions. We show that revision choice can affect the analysis of historical trade data. Second, we propose basic rules for translating historical, unstandardized trade statistics to the SITC. Third, we translate German product-level trade data from the first globalization to both SITC revisions 2 and 4 in order to find out which revision may be more applicable to historical data. Fourth, we then develop metrics to quantitatively assess our translation exercise. We argue that despite inevitable imperfections, applying the SITC yields useful results, even on a very disaggregated level.

摘要

我们研究标准国际贸易分类（SITC）。成千上万的研究依赖于分类的贸易数据，但是很少研究这些研究的分析单元（按特定层次排列的商品类别组合）的质量。通常不清楚产品或品种到底是什么。同时，从1950年代开始，越来越精细的贸易数据被从需要标准化的档案中取出。SITC为此提供了一个框架。我们做出四点贡献：首先，我们确定每个SITC修订版的特殊性，分析修订版之间的关系，并提供所有修订版之间的改进对应表。我们表明修订版本选择会影响对历史贸易数据的分析。第二，我们提出了将历史的，非标准化的贸易统计数据转换为SITC的基本规则。第三，我们将德国产品级别的贸易数据从第一次全球化转换为SITC版本2和4，以便确定哪个版本更适用于历史数据。第四，我们然后制定度量标准以定量评估我们的翻译工作。我们认为，尽管存在不可避免的缺陷，但即使在非常细分的水平上，应用SITC也会产生有用的结果。