The effect of regulatory benchmarks on firm reporting behavior

Yun Fan, Wayne B. Thomas, Chong Wang

Research output: Journal article publicationJournal articleAcademic researchpeer-review

5 Citations (Scopus)


Accounting researchers question the effect of (‘‘bright-line’’) regulations on firms’ financial reporting behavior. To offer evidence on this issue, we take advantage of the regulatory environment in the Chinese setting. In China, regulators have created a bright-line at which firms must disclose within 30 days after the fiscal year-end expected changes in net income of 50 percent or greater. Consistent with earnings management to avoid reporting earnings that miss this explicit regulation-generated benchmark, we find that far more firms than would be expected just beat the-50 percent threshold (i.e., there is a distinct ‘‘kink’’ in the earnings change distribution at -50 percent). As further evidence of earnings management, firms just beating the -50 percent threshold have higher abnormal accruals and excess non-operating income. The ability of firms to avoid missing the -50 percent regulatory benchmark, however, is reduced for those with stronger monitoring (foreign investors, exchange regulators, and IFRS). We find no evidence of earnings management at the regulatory threshold of +50 percent change in net income. Overall, our study sheds light on firms’ opportunistic reporting behavior to suppress bad news related to a regulatory benchmark and on the monitoring mechanisms associated with such behavior.
Original languageEnglish
Pages (from-to)85-107
Number of pages23
JournalJournal of International Accounting Research
Issue number1
Publication statusPublished - 1 Jan 2015
Externally publishedYes


  • Disclosure regulation
  • Earnings management
  • Monitoring
  • Regulatory benchmark

ASJC Scopus subject areas

  • Business and International Management
  • Accounting


Dive into the research topics of 'The effect of regulatory benchmarks on firm reporting behavior'. Together they form a unique fingerprint.

Cite this