Tunnelling, Fraudulent Financial Statements and Regulation Effects: Chinese Evidence

ABSTRACT

This study investigates the impact of regulatory intervention on tunneling through inter-corporate loans in Chinese fraudulent firms. We find fraudulent firms have significantly higher tunneling through inter-corporate loans than matching firms, and suggest that controlling shareholders are motivated to delay the recognition of the ensuing loss in the financial statements to make it more difficult for auditors to detect tunneling. In 2006, new CSRC regulation introduced responsibility by board chairman to resolve tunneling issues while the two Chinese stock exchanges initiated a ‘name and shame’ of tunneling individuals and amounts tunneled through public media. We find that tunneling balances in a sample of fraudulent firms reduce significantly after the announcement of this strict anti-tunneling regulation. Relative to a matched sample, fraudulent firms experienced improved operating performance and higher abnormal returns following the imposition of the new anti-tunneling regulations. We also find that informal institutions, such as social trust play an important role in mitigating tunneling. Overall, our results suggest the means by which Chinese regulatory mechanisms have been effective in protecting minority shareholders from expropriation by controlling shareholders.

KEYWORDS:

• Tunneling
• fraudulent financial statements
• Chinese listed firms

JEL:

• G30
• G32
• G38
• K22

Acknowledgments

We would like to thank Paresh Kumar Narayan (the editor) and two anonymous reviewers for their helpful comments and suggestions.

Supplemental Material

Supplemental data for this article can be accessed on the publisher’s website.

Notes

1. In 2006, the CSRC noted the pervasiveness of tunneling practices through excess borrowing by controlling shareholders, and officially labeled it “non-operational fund occupancy”.

2. The 2006 regulatory year observations are included in the analysis, but the results remain materially the same when they are excluded.

3. Due to the inherent stability of institutions, social trust can remain stable for a long period of time. We follow Wu, Firth, and Rui (2014) and Zhang and Ke (2002), using the nationwide survey conducted by the “Chinese Enterprises Survey System” in 2000 to estimate the level of social trust in the province where the company is headquartered.

4. Untabulated results show that median differences between fraudulent firms and matching firms are also similar.

5. The VIF for each variable is no greater than 4; hence, multicollinearity is not a problem.

6. In our un-tabulated OLS regression results with industry and year dummies, the coefficients on FRAUD are 0.038 in the low social trust subsample and 0.080 in the high social trust subsample. Both are significant at the 1% level. The difference is also significant at the conventional level, which supports our argument that greater tunneling balances differences between fraudulent firms and matching firms occur in the high social trust provinces.

7. Similar (un-tabulated) results are also found when we industry-adjust ROA, return on equity and return on sales by subtracting the industry median to allow for potential macro-economic impacts (Tu and Yu 2014).