On the reliability (or otherwise) of SIC codes

Researchers often make use of SIC (Standard Industrial Classification) codes when gathering and analysing data about the activities of companies. The use of these codes is, however, fraught with potential difficulties, errors easily creeping in and consequently distorting results. This paper outlines the major SIC code systems in use on both sides of the Atlantic – which, despite efforts to standardise them (and thus to make them worthy of their name), still present levels of inconsistency and unreliability both internally and comparatively – and discusses various problems associated with using the codes as data sources.