Social comparison of cost behaviour and financial analysts

Madadian Oveis, Walter Aerts, Tom van Caneghem

    Research output: Contribution to journalArticleScientificpeer-review

    Abstract

    We investigate whether social comparison of a firm’s reported selling, general and administrative (SG&A) expenses affects financial analysts’ information uncertainty (and their behaviour). Based on a sample of US firms, we examine whether similarity of a firm’s SG&A to an industry-specific peer-based benchmark (or social benchmark) is associated with analyst forecast dispersion, forecast error and coverage. For external observers, the SG&A relative to sales (SG&A ratio) is a key diagnostic of a firm’s cost behaviour, but interpretational ambiguity of the SG&A signal is likely to incentivise search for information-relevant external cues to set expectations about and assess a firm’s SG&A ratio. Higher similarity to the social benchmark is expected to attenuate information asymmetry between analysts and firms regarding firms’ ability to effectively control overheads, decreasing analyst information uncertainty about cost behaviour and performance. In line with a varying weights model for social comparison, we observe a negative association between SG&A similarity and both forecast dispersion and error of one-year-ahead earnings for firms with a prior SG&A ratio exceeding the social benchmark. Our findings also show a negative relationship between SG&A similarity and analyst coverage, especially for firms with a prior SG&A ratio exceeding the social benchmark.
    Original languageEnglish
    Pages (from-to)805-839
    JournalAccounting and Business Research
    Volume48
    Issue number7
    DOIs
    Publication statusPublished - Feb 2018

    Keywords

    • SG&A
    • social comparison
    • analyst forecast properties
    • peer-based benchmarking
    • aspirations

    Fingerprint

    Dive into the research topics of 'Social comparison of cost behaviour and financial analysts'. Together they form a unique fingerprint.

    Cite this