The Effects of Non-Audit Services and Audit-Firm Tenure on Earnings Conservatism: Evidence from Family Firms in Malaysia
35 Pages Posted: 11 Jan 2010 Last revised: 3 May 2012
Date Written: April 15, 2010
This paper examines the effects of non-audit services and audit-firm tenure on earnings conservatism in family firms. We hypothesize that in family firms, where there is no clear separation between managers and owners, auditor independence is more likely to be impaired when the incumbent auditor supplies a high volume of non-audit services or has either a too short or too long audit tenure. We expect that the impairment of auditor independence would result in lower earnings conservatism as the auditors acquiesce to management opportunistic reporting. Employing the asymmetric timeliness of earnings model introduced by Basu (1997), we find that in family firms a high provision of non-audit services and short audit-firm tenure (less than four years), are both associated with low earnings conservatism. For non-family firms, neither audit-firm tenure nor the provision of non-audit services have a significant relationship with earnings conservatism, implying no evidence of impairment of auditor independence in non-family firms. A test of the joint effect of audit-firm tenure and the provision of non-audit services in the family firm sample shows that firms with short audit-firm tenure and high non-audit fees have lower earnings conservatism, but no significant joint effect is found in the non-family sample. Our results are robust even after conducting several specification tests, including employing variation in the controls, using a non-price earnings conservatism measure, and alternative definitions of family firms, non-audit services and short audit-firm tenure. These findings provide valuable evidence from a developing country on the debate surrounding the impact of the provision of non-audit services and audit-firm tenure.
Keywords: Earnings Conservatism, Auditor Tenure, Non-audit services
JEL Classification: M41
Suggested Citation: Suggested Citation