Does financial statement comparability mitigate corporate frauds in an emerging market? Evidence from China

ABSTRACT

In this paper, we empirically examine whether financial statement comparability mitigates corporate fraud in China. Using the FSC measure proposed by De Franco, Kothari and Verdi (2011), we find that firms with greater comparability are less likely to commit frauds, either accounting – or non-accounting-related frauds. Further tests confirm that regulators can more quickly detect the fraudulent activities of accused firms if their financial statements are more comparable with those of their same-industry peers. Cross-sectional analyses show that the negative relationship between FSC and fraud incidence is more pronounced for firms with lower institutional ownership, and for those operating in regions with more developed markets. Overall, our study provides evidence for the benefits of peer comparisons in the fraud context, and has implications for investors, regulators, and standard setters.

KEYWORDS:

• Peer comparison
• financial statement comparability
• corporate fraud
• information asymmetry

Acknowledgments

Shiyang Hu acknowledges financial support from the National Natural Science Foundation of China (Grant No. 71802029). Yuanyuan Liu acknowledges financial support from the National Natural Science Foundation of China (Grant No. 71790602,72002182).

Notes

1. CASE – Basic Rules (2014) requires accounting information provided by enterprises to be comparable. Specifically, when recording the same or similar transactions or economic events, different enterprises should use the accounting principles allowed by the standards, to ensure that accounting information is comparable between entities (Rule 15).

2. Our research is also related to, but distinguishable from, prior literature investigating the relationship between FSC and management opportunistic behavior, including earnings management and restatement (Sohn 2016; Chen and Gong 2019), which differ from corporate fraud in terms of severity and frequency of misconduct. Using the DKV measure, Sohn (2016) finds that comparability is negatively associated with accrual-based earnings management (AEM) and positively associated with real earnings management (REM). This study highlights that comparability motivates firms to switch from AEM to REM, which does not suggest that comparability diminishes management opportunistic behavior. Chen and Gong (2019) also use the DKV measure and find that comparability is positively associated with financial reporting quality and managerial forecast accuracy and precision. Their main argument is that comparability improves the ability of managers to develop more accurate estimates and signal future firm performance, rather than constraining earnings management.

3. As emphasized by Yip and Young (2012), the characteristics of similarity and difference are equally important when examining comparability. Thus, comparability is conceptually related to but distinguishable from earnings similarities. Lang, Maffett, and Owens (2010) suggest that similarities in intra-industry earnings may be indicative of uniformity rather than true comparability. Besides, Choi et al. (2019) and Kim, Kim, and Musa (2018) document that comparability improves the information content of earnings, whereas Jackson, Li, and Morris (2020) find that earnings co-movements diminish earnings informativeness, which provides evidence of the difference between comparability and earnings co-movements. Thus, we do not use earnings co-movements as a proxy for comparability in the main tests.

4. In addition, the DKV measure is suitable in our approach to fraud detection. If fraudulent activities are recorded by an accounting system, they will eventually be reflected in the financial statements. Thus, outside investors and regulators are likely to identify unusual economic events from irregularities in comparable financial statements. Numerous recently published high-quality studies therefore use the DKV measure to test their main arguments (e.g. Choi et al. 2019; Jiu, Liu, and Liu 2020; Nam 2020).

5. CSRC and the two stock exchanges assign staff to regularly monitor companies in each specific industry. As companies in the same industry usually conduct similar economic activities, staff can develop industry expertise and conduct more effective supervision.

6. The industry classification follows the Industrial Classification Codes for Listed Companies (Edition 2012) issued by the CSRC.

7. We follow the sensitivity analysis (p.1361) of Jackson, Rountree, and Sivaramakrishnan (2017) to include both the DKV measure and earnings co-movement measure into the regression model. Unlike Jackson, Rountree, and Sivaramakrishnan (2017), who find DKV has little incremental effect in the U.S. setting, we find the coefficients on the DKV measure remain significantly negative. We also find that the coefficients on the co-movement measure are all negative but vary from significant to insignificant. The results suggest that the DKV measure is robust to the inclusion of earnings co-movement in our empirical tests. In other words, the coefficients on the DKV measure are more robust than earnings co-movement using a sample of Chinese listed companies. The institutional background and IFRS-convergent accounting standards of China may be the two possible explanations for the differences between Jackson, Rountree, and Sivaramakrishnan (2017) and our results. Further research is needed to explore and verify the potential reasons.

8. Results hold for Compmean and Compmedian (untabulated).

Additional information

Funding

This work was supported by the National Natural Science Foundation of China [71802029,72002182].