Weighted Crosswalks for NAICS and SIC Industry Codes

We provide weighted crosswalk files for the purpose of bridging Standard Industrial Classification (SIC) codes and North American Industry Classification System (NAICS) codes. SIC codes were the standard industry classification system for decades, but they eventually couldn’t keep up with the changing industrial structure of the 1990s. NAICS codes took over in 1997 and quickly became the reporting system for most government statistics in the US and elsewhere. This switch poses a problem for researchers since it imposes an artificial break in time series data. Unweighted crosswalk tables can help connect the two systems, but the many splits and merges create mappings that are not one-to-one. For situations that require one-to-one translations, researchers find themselves guessing as to the best match. And for situations where splitting and merging is acceptable, choosing weights sometimes feels arbitrary. The crosswalks included here contain weighting variables that make it possible to smoothly bridge between systems and construct consistent time series in a nonarbitrary way. Three different weighting schemes are included. The first based on employment, the second based on number of establishments, and the third based on total payroll.

本数据集提供加权匹配对照表文件，用于衔接标准行业分类（Standard Industrial Classification，SIC）代码与北美行业分类系统（North American Industry Classification System，NAICS）代码。SIC代码曾长期作为主流行业分类标准，但最终无法跟上20世纪90年代不断变化的产业结构。NAICS代码于1997年取代SIC，迅速成为美国及其他地区多数政府统计数据的报送分类体系。这一分类标准的更替给研究者带来了难题：它人为造成了时间序列数据的断点。非加权匹配对照表虽可用于衔接两类分类体系，但由于行业分类存在大量拆分与合并的情况，二者间的映射关系并非一一对应。当研究需要一一对应的映射关系时，研究者往往只能通过主观猜测选取最优匹配结果；而在允许分类拆分与合并的场景中，权重的选取往往带有主观随意性。本数据集附带的匹配对照表包含权重变量，可实现两类分类体系的无缝衔接，并以客观非主观的方式构建一致的时间序列数据。本次数据集共提供三种不同的权重设定方案：第一种以就业人数为权重依据，第二种以企业数量为权重依据，第三种以工资总额为权重依据。