Direct and Indirect Effects of Internal Control Weaknesses on Accrual Quality: Evidence from a Unique Canadian Regulatory Setting

Abstract:     
Public disclosure about effectiveness of internal control systems is subject to much controversy in Canada, resulting in Canadian disclosures being made in Management Discussion and Analysis (MD&A). These disclosures are provided to investors without a definition of the weaknesses to be reported, without implementation effectiveness testing, no direct management certification and no external audit of such disclosures. Though the cost of such SOX North disclosures are lower than in the U.S setting, the credibility of these disclosures is far from assured, posing an important empirical issue for regulators interested in the cost versus benefit trade-off of various disclosure regimes. We use OLS regression and path analysis embedded in a structural model to examine the association between the strength of internal control and accrual quality in order to infer whether these disclosures are credible. The OLS regression finding of an overall negative net effect of internal control weaknesses on accrual quality is refined by recursive path analysis which shows that this overall negative effect is composed of a large negative direct effect of weaknesses on accrual quality and a smaller offsetting positive indirect effect of weaknesses on accrual quality via increased audit effort. The modest substitution effect implies that auditors cannot fully compensate for poor internal control by increased substantive work, which is per se evidence justifying some form of internal control disclosures for investors. Overall, our results suggest that lower cost SOX North disclosures are credible. This informs the cost-benefit debate facing regulators around the globe seeking alternatives to the costly U.S. model of internal control reporting.

Number of Pages in PDF File: 53

Keywords: Sarbanes-Oxley, Internal control, Accrual quality, Material weakness

JEL Classification: M40, M41, M43, M46, G34, M49, G38

Accepted Paper Series