Leveraging XBRL Calculation Linkbases to Overcome Semantic Heterogeneity across XBRL Fillings: The Multi-Ontology Multi-Concept Matrix (M3)

In 2008, the US Securities and Exchange Commission (SEC) mandated that large accelerated filers issue financial statements in eXtensible Business Reporting Language (XBRL). One purported benefit of issuing financial statements in a machine readable language is the facilitation of automated inter firm comparisons. However, XBRL, an XML-based language, is extensible by each individual filer. This extensibility compromises the ability of an automated agent to make meaningful comparisons between firms on any given metric or financial concept by introducing semantic heterogeneity across filings. In this paper we make an account of the development of an automated method for the generation of ontologies for XBRL calculation linkbases. In addition, we demonstrate a novel method for resolving semantic heterogeneity which leverages the expressivity of XBRL calculation linkbases. Our major premise is that the meaning of a given financial concept is captured by its relative position in a calculation hierarchy. The representation of calculation linkbases using an ontology language thus becomes the basis of our efforts towards the resolution of semantic heterogeneity across XBRL filings in the US jurisdiction.