Stock price contagion effects through investment banks

ABSTRACT

This paper investigates how stock market investors react to non-fraudulent firms that share the same investment bank with fraudulent companies. Using a Chinese sample from the period of 2003 to 2018, we find that firms penalized for IPO or M&A fraud induce stock price declines among non-fraudulent firms which share the same investment banks (non-fraudulent contagion firms). The results also show that stock price declines are more pronounced for low-quality investment banks, and investors impose larger penalties on stock prices when non-fraudulent client firms are of lower earnings quality, weaker corporate governance, and higher information asymmetry. Furthermore, we demonstrate that the non-fraudulent contagion firms are more likely to commit accounting fraud and exhibit inferior long-term post-IPO/post-M&A performance. Overall, the findings indicate an important stock price contagion effect occurring at the investment bank level.

KEYWORDS:

• Stock price contagion effect
• Investment bank
• Information transfer
• Market reaction

JEL CLASSIFICATION:

• G24
• D80
• G12

Acknowledge

We acknowledge partial financial support from China’s Management Accounting Research and Development Center in Central University of Finance and Economics. Yanan Zhang acknowledges financial support from the National Natural Science Foundation of China (No. 71902206). Bing Zhu acknowledges financial support from the China National Natural Science Foundation (No.71802207) Humanity and Social Science and Youth Foundation of the Ministry of Education of China (18YJC630271) and Young Talents Development Support Scheme of Central University of Finance and Economics (QYP2004). All errors are our own.

Disclosure of potential conflicts of interest

No potential conflict of interest was reported by the author(s).

Notes

¹ In our study, we investigate how stock market investors react to non-fraudulent firms that share the same investment bank. As stock prices already contain all publicly available information, endogeneity is less of a concern in our paper.

² In this paper, we use sponsor and underwriter interchangeably.

³ The standards of MAR are as follows: (1) purchasing or selling assets that accounted for 50% or more of total assets in the last year's audited financial statements; or (2) purchasing or selling assets that generated revenue accounting for 50% or more of the total revenue in the most recent year's audited financial reports; or (3) purchasing or selling net assets that accounted for 50% or more of net assets in last year's audited financial reports, and more than 50 million RMB; or (4) not to the above-mentioned extent, but the CSRC finds out that there exist some big problems that may damage the legitimate rights and interests of the listed company or investors.

⁴ We manually check each of the sanction announcement dates. When there is more than one date related to the same fraud sanction, we employ the earliest one as the announcement date to calculate CARs in price contagion tests.

⁵ We use 120 trading days to estimate normal stock return and choose 11 days as time period between normal return estimation window and event window. On the one hand, this time period should be short to generate more accurate normal return; on the other hand, it could not be too short to be affected by the event. Therefore, following Boehmer, Masumeci, and Poulsen (1991), we use 120 trading days of return data ending 11 days before the M&A announcement. Furthermore, we perform robustness tests and use market model parameters which are estimated using 120 trading days of return data ending 31 days before the M&A announcement. Untabulated results show that the main results are still there.

⁶ Of all the investment banks that participated in the classification during 2010-2018, 152 (15.75%) were labelled as AA, 254 (26.32%) were labelled as A, 85 (8.81%) were labelled as B, 133 (13.78%) were labelled as BB, 248 (25.7%) were labelled as BBB, 14 (1.45%) were labelled as C, 25 (2.59%) were labelled as CC, 51 (5.28%) were labelled as CCC, and 3 (0.31%) were labelled as D. No companies were evaluated as AAA or E.

⁷ We notice that the coefficients of controls variables are largely insignificant and only the constant is significant. To address this issue, we perform the correlation tests and find that there are significant correlation between CARA(−1,+1), (CARA(−2,+2)) and Roa, Top1, Contagion_IB. However, after we control Year, Industry, and Investment bank fixed effects, the coefficients of these three variables are not significant any more. It suggests that the effects are absorbed by those fixed effects. Moreover, untabulated results show that the constant becomes insignificant after province fixed effect is included. Therefore, we think that province fixed effect may be the unobservable information not captured in the model.

Additional information

Funding

This work was supported by the National Natural Science Foundation of China [No. 71902206, No.71802207]; Humanity and Social Science and Youth Foundation of the Ministry of Education of China [18YJC630271]; Young Talents Development Support Scheme of Central University of Finance and Economics [QYP2004].