Media tone, bias, and stock price crash risk: evidence from China

ABSTRACT

This paper studies the effect of media tone on firm-specific price crash risk. Media tone measures the imbalance between positive and negative coverage. Using data from China, we find firms with more favorable media tone have higher crash risk. The transmission channel is through weakened media governance and the subsequent motivated managerial opportunism. Moreover, the effect is more prominent for firms having more advertising expenditures, institutional holdings, and analyst coverage. We also find only the tone of state-controlled media significantly affects crash risk. Our study suggests the existence of media bias and its economic consequence.

KEYWORDS:

• Media tone
• media bias
• price crash risk

Acknowledgments

We are very grateful to the associate editor Junbo Wang and the anonymous reviewers for their valuable comments on the paper.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Luo et al. (2016) define a firm’s political connections as its CEO or board chairperson having political experience, which includes serving in the government, the Party committee, the People’s Congress, the People’s Political Consultative Conference, the People’s Court, the People’s Procuratorate, or the People’s Bank at both local and central levels.

2. Piotroski, Wong, and Zhang (2015) focus on two major political events in China: meeting of the National Congress and provincial-level political promotions.

3. In 2014, one of the largest nation-wide business press, Chinese business news website 21cbh.com has had its license revoked and all staff have been laid off because it is reported to collaborate with two public relations firms to extort money from companies in return for favorable coverage and withholding negative news reports on the site. If companies refused, the website would publish negative or malicious information about the company. For more details, please refer to http://www.chinadaily.com.cn/china/2014-09/12/content_18591632.htm .

4. Risk-taking could be another cause of bad news holding because it makes the performance of firms worse. However, from an agency perspective, risk-taking behavior is not a self-interest behavior of management. Instead, some studies suggest that managers do not take a risk in order to prevent short-term earnings decline or losses, since risk-taking is often related to high-risk and long-term investments, such as R&D activities (Baber et al. 1996; Hirshleifer, Teoh, and Low 2012). Since the internal mechanism through which media tone may lead to a price crash is based on managerial opportunism, risk-taking is not the internal mechanism in this paper.

5. A shares are shares that denominated in Chinese Yuan (RMB) and traded in the Shanghai and Shenzhen stock exchanges.

6. An investment banker may use media to market an IPO to sentiment investors (Cook, Kieschnick, and Ness 2006), which induces significantly positive media tone during IPO events, so we delete the samples listed within 3 years.

7. The disclosure requirements and accounting rules of the financial services industry are significantly different from other industries in Chinese stock market, so we delete the samples in this industry.

8. In the stock market of China, a firm under special treatment (with ‘ST’ or ‘*ST’ in stock name) means that the firm has a risk of delisting. Since this kind of firms continuously have serious problems in operation, the media tone may be significantly negative.

9. There may be endogeneity between media tone and overinvestment. That is, the overinvestment of firms makes their performance worse, then the firm manages the media to issue positive reports to cover their activities. To address this, we impose a one-year lag between Tone1 and Overinvestment to test whether media tone in year t can predict the overinvestment in year t + 1. The regression result shows that media tone will increase the future overinvestment of companies. The results are shown in Table A2 in Appendix A. Moreover, we use Panel VAR and Granger Causality Test to show the relation between media tone and overinvestment. The result shows that media tone does Granger Cause overinvestment, whereas overinvestment does not Granger Cause media tone, suggesting that media tone does have the ability to predict overinvestment. The results of Granger Causality Test are shown in Table A3 in Appendix A. Since the Granger results may be driven by omitted variables, we also use the industry average media tone and the location average media tone as instrumental variables to mitigate this concern. The 2SLS results show that the instrumented variable (predicted Tone1) are positive and significant. The corresponding results are shown in Table A4 in Appendix A. Collectively, the results from these tests indicate that endogeneity concerns are not likely to be driving our evidence.

Additional information

Funding

This work is supported by the National Social Science Foundation of China [Grant No. 18XJY024].