Do companies disclose intellectual capital in their annual reports? New evidence from explorative content analysis

Abstract

PurposeThis study suggests an alternative to confirmatory content analysis (CA) and empirically demonstrates that explorative CA enables new insights into the mechanism of intellectual capital (IC) disclosure. In so doing, this research contributes to both methodological and empirical advancements in IC disclosure research.Design/methodology/approachEmploying the assumptions of positive accounting theory and taking book value of intangible assets as a reference, our research design utilizes well-established text-mining (TM) tools based on a least absolute shrinkage and selection operator regression. We assume that the degree of cohesion between officially disclosed and evaluated intangible assets on balance sheets and those contextually delivered in narrative form may affect how IC is ultimately disclosed in annual reports.FindingsOur main finding is in line with the results and criticism of previous studies. We show that companies do not extensively disclose IC in their annual reports. However, some narrative forms for IC disclosure are identified and confirmed by several robustness checks.Research limitations/implicationsFirst, the findings provide internal validity only for large US enterprises. These firms have similar, well-structured reporting requirements. This analysis might be enriched by an examination and a comparison of different institutional contexts, such as emerging countries. Second, following previous studies, annual reports serve as the source of data. Consequently, the findings are relevant only for mandatory and voluntary disclosure of IC, mitigating the relevance of this study for contexts of involuntary disclosure.Originality/valueThis study makes two contributions. First, we add to the empirical literature by offering one more piece of evidence on whether and, if so, the extent to which companies disclose IC in their annual reports. Second, we provide further examination of confirmatory CA by proposing a number of statistically validated codes and tokens that are indicators of IC communication by companies.