Can financial analysts constrain real earnings management in emerging markets? Evidence from China

ABSTRACT

We examine the effect of financial analysts on real earnings management in China. Our empirical results show a negative relationship between analyst coverage and real earnings management. This relationship exists within the full sample and sub-sample of firms meeting or beating earnings benchmarks. It survives tests to address endogeneity problems. We also find that real earnings management can impair companies’ profitability and analysts will revise their earnings forecasts downward for firms with aggressive real earnings management. Further research suggests that analysts can also curb accrual earnings management.

KEYWORDS:

• Analyst coverage
• real earnings management
• profitability
• forecast revisions

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Examples include the Interim Regulation over the Management of Securities and Futures Investment Consultation (China Securities Regulatory Commission, 1997), Notice on Some Issues Concerning Standardizing the Conduct of Public Securities Investment Advisory Business (China Securities Regulatory Commission, 2001), Interim Provisions on Securities Investment Consulting Business (China Securities Regulatory Commission, 2010), and Code of Conduct for Securities Analysts (Securities Association of China, 2012). These can be found at http://www.csrc.gov.cn/pub/newsite/flb/flfg/, http://www.sac.net.cn/flgz/zlgz/and http://www.csrc.gov.cn/pub/newsite/flb/flfg/bmgf/zjfw/tzzx/201012/t20101231_189937.html.

2. We do not combine PRODEM, CFOEM and DISXEM as an aggregate measurement because the same activities lead to abnormally high production costs also lead to abnormally low CFO; thus adding these two amounts results in double counting. We thank the referee for pointing this out to us.

3. (3.178-1.099)×0.023=0.048.

4. For brevity, we suppress the results of control variables in this and the left tables.

5. We thank the referee again for pointing out that earnings surprise should be controlled.

6. For brevity, only regressions results of RM1and RM2 are presented in the table. The results for other three individual proxies are similar.

Additional information

Funding

This work was funded by the National Natural Science Foundation of China under Grant [71572054] and the Hunan Province Social Science Foundation under Grant 2018YBA087.