Open AccessThe textual similarity of KAM disclosures for Spanish companies
Author(s) -
Sheng-Feng Hsieh,
Cleber Beretta Custodio,
Miklos A. Vasarhelyi
Publication year - 2021
Publication title -
the international journal of digital accounting research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.288
H-Index - 13
eISSN - 2340-5058
pISSN - 1577-8517
DOI - 10.4192/1577-8517-v21_7
Subject(s) - similarity (geometry) , audit , cosine similarity , accounting , business , computer science , artificial intelligence , cluster analysis , image (mathematics)
We investigate and document the textual similarity of key audit matter (KAM) disclosures by using KAM items in auditor’s reports of Spanish companies in fiscal years 2017 and 2018. The main objective is to understand how similar KAMs are disclosed from one year to another. Following prior literature, we use the cosine similarity to measure the textual similarity between KAM items in terms of word usage. We classify and analyze KAM items for two consecutive years based on the following three combinations: (1) KAM topic, (2) KAM topic and auditor, and (3) KAM topic, auditor, and industry of the client being audited. The results indicate that auditors from the same accounting firm tend to have a recurring textual similarity under each KAM topic, and such similarity increases for clients within the same industry. The results add empirical evidence to the understanding of the recurring textual similarity of KAM disclosures