Does Board Independence Affect Audit Fees? Evidence from Recent Regulatory Reforms

Abstract

To enhance board oversight, since 2002, US legislation has required listed companies to have a majority independent board. This paper uses this legislative change to examine the relation between board independence and audit fees. To provide a clean estimate of this relation, we adopt a difference-in-difference approach using a sample matched on client firm characteristics. We find that greater board independence is insignificantly associated with a change in audit fees when client firms operate in a weak information environment. When the information environment is strong, greater board independence is associated with an increase in audit fees. Our results are consistent with the nascent theory emphasizing information asymmetry and provide insight into the effectiveness of the mandated board independence in relation to audit quality.

Acknowledgements

We are grateful for the helpful comments from Laurence van Lent (editor), an anonymous referee, Kevin C. W. Chen, Neil Fargher, Jeong-Bon Kim, Pauline Weetman and workshop participants at the University of Edinburgh, and City University of Hong Kong.

Supplemental Data and Research Materials

Supplemental data for this article can be accessed on the Taylor & Francis website, doi: 10.1080/09638180.2015.1117007.

Notes

1 See SEC press release 2002–2023, 13 February 2002.

2 Since endogeneity is often caused by unobservable firm characteristics and omitted variables, there are advantages in using a difference-in-difference method with a sample matched on firm characteristics to mitigate these problems (Roberts & Whited, 2013). However, there are limitations to what econometric methods can achieve. Our design, combined with the setting of a significant legislative change, can, to some degree, address the endogeneity embedded in prior research, but our design does not intend to offer a definitive treatment of the issue regarding the relation between board independence and audit fees and unobservable heterogeneity can still not be ruled out.

3 The economic magnitude of this increase is about $200,000, on average, which is due to 20% (after unclogging) of the sample mean audit fees (about $1 million). This result is close to what is reported in the literature (i.e. Carcello et al., 2002).

4 Specifically, an independent director cannot be an employee or a family member of an executive officer of the company. An independent director cannot receive more than $120,000 in compensation from the company, other than director and committee fees (NYSE 303A.02).

5 NYSE 303A.07. SOX, which became effective on 30 July 2002, has similar requirements.

6 See note 1.

7 Since audit fees are determined through fee negotiations, either a more powerful audit committee or more powerful management (CFOs) can exercise bargaining power on behalf of the organization (Doty, 2011; Hellman, 2011). Despite the fact that SOX makes the audit committee directly responsible for determining auditor compensation, management (CFOs) still exerts significant influence in fee negotiations (Cohen et al., 2010).

8 Propensity score matching is considered superior to the Heckman (1979) selection models, since it provides a more direct estimate of the treatment effects. Propensity score matching is an appropriate design choice for our study, since the attribute-based matching will naturally phase out the effects of observable differences in compliant and non-compliant firms’ characteristics on their audit fees. Using a logit model is the most common method to estimate propensity scores (Guo & Fraser, 2010).

9 We include the natural log of the closing share price (SP) to control for market valuation and stock liquidity of the firm (Chordia, Huh, & Subrahmanyam, 2006; Utama & Cready, 1997), which is thought to be relevant to the model of board independence (Armstrong et al., 2014).

10 As stated above, the new rule of majority independent boards was introduced by stock exchanges (NYSE and NASDAQ) in 2002 and firms were required to comply with it by no later than the end of 2004 (i.e. 2002 and 2003 are transitional years). Therefore, 2000 is selected as the pre-regulation benchmark year and 2005 as the post-regulation benchmark year.

11 Around this time period, there was a general improvement in the information environment due to the introduction of SOX and related regulatory changes. To address potential endogeneity between board independence and information environment, we measure HIX in 2000 (i.e. in the pre-regulation period) to make sure that HIX is not confounded by the change in board independence from the pre- to the post-regulation period (Chen et al., 2015; Duchin et al., 2010).

12 See http://www.rhsmith.umd.edu/faculty/sbrown.

13 In the untabulated results, we also include HIX besides the interaction of NC in the regression. The results are qualitatively similar to those reported in the paper. The coefficient on HIX is negative, suggesting firms with a stronger information environment before the regulatory change experience a smaller increase in audit fees after the regulatory change. Since there is a general improvement in information environment with the regulatory changes (such as SOX and the new listing rules) around this time period, this result is consistent with the assumption that the relation between information environment and audit fees is positive.