External auditor types and the cost stickiness of listed companies

Abstract

External auditing is an important method of corporate governance that can effectively monitor and restrict a manager’s opportunistic behaviour and reduce cost stickiness. We focus on Chinese A-shares listed companies from 2002 to 2010 to explore the governance role of external auditors in cost stickiness. We classify audit firms into three types: the Big Four International, the Big Ten Domestic, and the non-Big Ten Domestic. The results show that the cost stickiness of firms audited by the Big Four International is significantly lower, while the cost stickiness difference between firms audited by the latter two types is not significant. Further tests show that the restriction on cost stickiness for the Big Four International is more significant when the largest shareholder’s shareholding ratio is lower or when the local market is less developed. These results indicate that the restriction placed by audit quality on cost stickiness varies with other corporate governance mechanisms. Our results are qualitatively unchanged under different robustness tests, including propensity score matching. This study shows that Big Four International audit firms can better reduce cost stickiness, which provides evidence of the higher audit quality of these firms and extends the study of restrictive factors on cost stickiness to external corporate governance mechanisms.

Keywords:

• Big Four International
• cost stickiness
• shareholding ratio
• marketisation level
• propensity score matching

View correction statement:

Erratum

Acknowledgements

This research is supported by grants from the National Natural Science Foundation of China (No. 71402198), the Beijing Municipal Commission of Education ‘Joint Construction Project’, the Beijing Municipal Commission of Education ‘Pilot Reform of Accounting Discipline Clustering’, ‘2011 Synergetic Innovation’ Key Project on ‘Development of Public Accounting Profession’ and ‘121 Talent Project for Young Doctoral Graduates Development 2014’ of the Central University of Finance and Economics.

Notes

1. Chen, Lu, and Sougiannis (2012) provide direct evidence for this argument. They find that cost stickiness has a positive correlation with corporate free cash flow and this positive correlation can be weakened when the corporate governance level is higher. Calleja et al. (2006) find that expense stickiness in France and Germany is higher than in the United States or Britain in their international study and they conjecture that the difference in corporate governance is the main reason.

2. China is a large country where the levels of market development and government intervention vary significantly across the 30 areas (provinces, autonomous regions or municipalities) (Jian & Wong, 2010). Fan, Wang, and Zhu (2010) offer an index that measures the development of marketisation of different areas in China.

3. In different years, it can be the Big Five International or the Big Eight International instead of Big Four International. For the sake of brevity, we use the term Big Four International for all conditions without differentiation. All these Big N international audit firms refer to the well-recognised top international audit firms.

4. The findings of these studies conflict for the following reasons, in addition to the difference in research focus (discretionary accruals, audit opinion types, ERC (earnings response coefficient), and conservatism): first, the sample period is short. Most sample periods are one year to three years and the longest are five years. The shorter the sample period, the more the variables of interest can fluctuate, especially discretionary accruals. Second, the difference in the sample size of firms audited by Big Four International and non-Big Four International is not well addressed. Third, the self-selection problem is paid little attention. Our paper, instead, chooses a nine-year sample period (from 2002 to 2010) to conduct the empirical tests and methods such as propensity score matching (PSM) to mitigate the potential problem.

5. The SME started in 2004 and GEM started in 2009. But the government stopped the IPO approval during the non-tradable share reform (2005–2006), so companies listed before 2007 are mainly on the main board.

6. Other reasons include the consideration of cost adjustment and managers’ optimism (Banker et al., 2011). Sun and Liu (2004) attribute these reasons to an opportunistic behaviour perspective, a contract perspective, and an efficiency perspective.

7. However, their research design does not control for common variables, economic factor variables, or year or industry fixed effects.

8. Calleja et al. (2006) mention external supervision theoretically, but do not conduct empirical research.

9. Such as overextending the number of employees (Williamson, 1963) and building luxurious offices.

10. Controlling earnings management is not the only way external governance works. If a company’s performance worsens and managers still spend 1 million RMB to build luxurious offices, they can cover the adverse influence of such expenses on performance through earnings management. A capable and independent auditor can either correct the earnings management during the auditing procedure, which reduces the degree of earnings management, or alert shareholders by signing non-standard audit opinions without correcting for earnings managements (in such cases, the degree of earnings management is unchanged).

11. Although many variables have been used in prior studies to measure agency costs, such as the administrative expense rate and free cash flow, most of these variables might present the consequence of the whole corporate governance including auditing (Gul & Tsui, 1998; Li, 2007).

12. Our sample starts in 2002 because the concept of the Big Four International tends to have become stable with the voluntary surrender by Arthur Andersen of its licence to practice in the US. Also, the Chinese CPA Association started issuing its Comprehensive Evaluation of the Top 100 Accounting Firms in 2002, which provides a good reference for defining the Big Ten Domestic audit firms.

13. The firm ranking is derived from yearly Comprehensive Evaluation of the Top 100 Accounting Firms released by China CPA Association (http://www.cicpa.org.cn/Column/swszhpm).

14. These differences further support the necessity of using PSM as a robustness test.

15. Some later tests show that the sum of the coefficients of ∆Income*Income_decrease and ∆Income*Income_decrease*Big4 is significantly positive, meaning that there is an anti-stickiness phenomenon. This kind of result is mainly found in groups with lower largest shareholder’s shareholding ratio and groups with a lower market index. We conjecture that one possible reason is that the agency cost in this kind of company is higher. When sales decrease, managers trying to achieve performance goals may have to adopt real earnings management if their audit firms (Big Four International) effectively restrict them from conducting earnings management. In such a situation, excessive cost cutting can lead to an anti-stickiness phenomenon. Cohen and Zarowin (2010) and Zang (2012) both point out that the degree of real earnings management is higher when firms are audited by the Big Eight International. We do not find an anti-stickiness phenomenon when PSM methods are used.

16. The medians of the largest shareholder’s holding ratio (Sh) and the market index (Marketisation_index) are 0.357 and 7.92, respectively.

17. We also estimate the model using 2×2 subsamples based on whether the firm is audited by Big Four International firms and whether the largest shareholder’s holding ratio (or market index) is higher than the median. The results show that the cost stickiness in groups when firms are audited by non-Big Four International firms and when the largest shareholder’s holding ratio (or when the market index is lower than the median) is the strongest.

18. The result is more significant if we compare the Big Four International sample with the non-Big Ten Domestic sample (e.g. the control group does not include the Big Ten Domestic sample).

19. The idea is to mitigate the influence of earnings management in the form of a big bath, instead of the influence of the absolute value of earnings management. In accordance with this idea, we control for DA instead of its absolute value. In robustness tests, we control for the absolute value of discretionary accruals (AbsDA) and the interaction term ∆Income*Income_decrease*AbsDA.

20. We think the restrictions on cost stickiness and earnings management are two parallel consequences. Big Four International firms can use different methods, such as restricting earnings management, signing non-standard audit opinions, and providing comparable financial information of higher quality, to help shareholders make executive compensation and turnover decisions to reduce managers’ opportunistic behaviour. Reducing earnings management is only one of these methods and, when the latter two methods are used, the degree of earnings management cannot completely explain the change in cost stickiness.

21. Unreported results are available from the authors upon request.

22. To address this problem, we also conduct subsample tests, but this method is convincing only in the test of the basic hypothesis. Further tests still rely on the basic model and may still be affected by potential model misspecification (the comparison between subsamples with different ownership structure or different external environments is a Chow test of the cubic item).