Auditor independence and non-audit services - ICAEW

CHAPTER 10EVIDENCE OF ASSOCIATION BETWEEN NAS FEES AND EARNINGSQUALITY10.1 IntroductionUntil very recently, there were no studies that examined whether the provision ofNAS impacts upon the attributes of accounting numbers. The absence of publishedNAS fee data in the US until recently is one reason for this. In the last two years,several (mostly as yet unpublished) papers have explored the link between earningsquality (or earnings conservatism or earnings management) and NAS. Earningsquality is, in part, a function of auditor independence.10.2 Earnings management and the measurement problemThe academic and professional literature offers a variety of definitions of earningsmanagement. A typical recent offering is ‘earnings management occurs whenmanagers use judgment in financial report and in structuring transactions to alterfinancial reports to either mislead some stakeholders about the underlying economicperformance of the company or to influence contractual outcomes that depend onreported accounting numbers’ (Healy and Wahlen, 1999, p.368).The earnings figure emerges from the application of accrual accounting, by applyingaccounting principles for revenue recognition and the matching of expenses withrevenue. The main reason for reporting earnings figures (in addition to cash flows) isto help investors assess the performance of the company, generally by smoothingreported earnings. However, given managerial incentives and the existence ofdiscretion in applying accounting principles, there is concern that companies are usingsuch practices to ‘stash accruals in cookie jars during good times and reach into themwhen needed in the bad times’ (Levitt, 1998). The key question, posed by (Dechowand Skinner, 2000) is: when does the appropriate exercise of managerial discretionbecome earnings management?Studies of this type require a measure of earnings management activity and the mostcommon approach is to use a measure of discretionary accruals 26 as a proxy for this.Discretionary accruals are measured using the Jones (1991) model or a variantthereof. To estimate discretionary accruals, studies first identify total accruals,measured as the difference between reported net income and cash flows fromoperations. Total accruals is then regressed on variables (such as revenues and grossfixed assets) that would explain the normal level of accruals. Discretionary accrualsthen emerge as the residuals from the regression model.10.3 Empirical studies10.3.1 US studiesChung and Kallapur (2001) make use of the recent NAS fee disclosure requirementsto investigate the link between client importance and abnormal accruals (estimatedusing the modified Jones model). Client importance is measured as the ratio of clientfees to audit firm’s total US revenues and also as the ratio of client non-audit fees toaudit firm’s total US revenues. After controlling for other variables that influence26 Also termed unexpected or abnormal accruals.56