| Summary: | 博士 === 逢甲大學 === 商學研究所 === 100 === This study first examines the association between the occurrence of accounting restatement and accrual-based earnings management. Using TWSE/GTSM-listed companies that had already announced restatements from 2001 to 2011 as the restating sample and further partitioning the restating sample into two subgroups (i.e., forced-restating firms and voluntarily restating firms), the paper performs logistic models to examine the association between restatements and accruals. The results indicate that firms announcing to restate their financial statements are more likely to engage in accrual management not only in the restating year, but also in the year prior to restatements than non-restating firms, and accruals of forced-restating firms in the year prior to restatements have the largest impact on the incident of restatements. In addition, restating firms engage in more accrual management from their operating activities and financing activities than non-restating firms, and forced-restating firms make more use of working capital accruals to engage in earnings management than voluntarily restating firms.
Second, this study investigates the association between real earnings management activities and the incidence of restatements, following Roychowdhury’s (2006) method. The results indicate that firms with restatements are likely to have more deficient cash flows from operating activities and frugal discretionary expenditures than non-restating firms, while firms with forced restatements are likely to have more deficient cash flows from operating activities and overproduction than firms that voluntarily issue restatements.
Third, the paper uses two-stage least squares (2SLS) models to explore market reaction to the announcement of restatements. Announcing restatements is highly likely to negatively influence stock returns in the time window (-1, +1), whereas no significant association occurs in the time window (-1, +3). Market reaction to restatements may have become less negative in recent years.
Finally, it is hypothesized that disciplining misstatements for a restating firm may result in the high likelihood of CEO turnover, CFO turnover, and auditor switch. The empirical results demonstrate that, although the incident of restatements appears to have no significant association with CEO turnover, the likelihood of CEO turnover is enhanced when CFO turnover follows the restatements and the restatements are involved in revenue recognition and inflated income. Meanwhile, CFO turnover has a direct and positive association with the incident of restatements, especially when CEOs also resign from their position in the same year as the restatement and the restatement pertains to revenues recognition and inflated income. Auditor switch shows no predicted association with the incident of restatements, but changes in the CEOs or CFOs during the restating year have the largest impact on auditor switch.
|
|---|
