Reexamining the influence of large clients on office-level auditor reporting decisions : A replication study

Purpose – This paper revisits the Reynolds and Francis’ (2001) study via the use of a more current dataset, incorporation of improvements into the accrual model and the use of actual fee data vs estimates. Using the improved analyses, the purpose of this paper is to examine whether more conservative auditors’ reports on larger clients are still evident. Design/methodology/approach – The paper follows Reynolds and Francis (2001) in using a regression model with White-adjusted t -statistics for the discretionary accrual model and a logistic model for going concern analysis. The most current discretionary accrual model is used to improve the original model, use actual fee data (not available previously), and add analyses using the two components of total fees (i.e. audit and non-audit fees). Findings – As opposed to Reynolds and Francis (2001), the results show that the Big Five auditors are less conservative with higher-paying clients as they allow their clients to have more discretionary accruals. While Reynolds and Francis (2001) found that auditors are more likely to report going concern opinions for higher-paying clients, the results in this paper does not show any difference in the propensity of auditors to issue going concern opinions. Originality/value – This study replicates Reynolds and Francis (2001) using more recent US data, applying the most recent discretionary accrual model, using the actual fee data, and adding analyses using total fees decomposition.