Can machine learning aid auditors’ judgment about goodwill impairment?

Purpose This paper aims to examine whether machine learning (ML) models can become a decision aid for auditors to improve their judgment about the occurrence of goodwill impairment and the value of goodwill impairment. Design/methodology/approach The authors used firm-year data from companies listed in the USA from 2002 to 2023. They use classical ML methods for both regression and classification modeling to predict firm-level goodwill impairments across two related data sets. Findings The results suggest that auditors can potentially use ML in predicting goodwill impairment, especially when making a judgment about the occurrence of goodwill impairment. Originality/value Goodwill impairment is an important accounting estimate that contributes to high audit risk. To the best of the authors’ knowledge, this study is among the first to apply ML models to assist auditors with judgment about goodwill impairment.