87 / 322
臺大管理論叢
第
26
卷第
3
期
87
Following Zhang (2007), this study used the growth rate of the number of employees as
a proxy for the magnitude of working capital management. Based on this measure, we
classified the sample companies into four groups. Companies with the highest and lowest
levels of the growth rate were classified as companies with a high magnitude of working
capital management. We compared the discretionary accruals in the groups with a high
magnitude of working capital management to those in the other groups and determined the
effect of working capital management on discretionary accruals. In addition, we performed
simulation tests according to procedures similar to those of Kothari et al. (2005) and
evaluated the Type I Error rate for various measures of discretionary accruals. Finally, we
analyzed particular samples, such as companies that increase capital (seasoned equity
offering) and companies that reduce capital (for the settlement of accumulated losses). The
analysis of these particular samples provides a different observation; that is, by controlling
working capital management, we reexamined the assertion that the management of firms that
increase capital (reduce capital) manipulates earnings upward (downward), as reported in the
literature.
3. Findings
The empirical results indicated that working capital management is positively related to
total accruals and discretionary accruals. In addition, companies with the highest working
capital management had the highest discretionary accruals, whereas the opposite applied to
companies in the group with the smallest working capital management. In addition, for the
two groups, our simulation tests for the null hypothesis of zero discretionary accruals
revealed that the phenomenon of over-rejection, indicating an estimation bias in the
discretionary accruals of the sample firms.
The empirical results for the sample companies that increased capital showed that
companies with a higher degree of working capital management had a higher level of
discretionary accruals than did those with a lower degree of working capital management.
Regarding the companies that reduced capital, this study determined that the discretionary
accruals of the companies with a higher degree of reduction in working capital management
were lower than those of the other companies. The estimation bias in the discretionary
accruals affected the assessment of whether firms issuing seasoned equity offerings or
reducing capital to settle accumulated losses manipulated earnings. In addition, by using
companies that increased capital as the sample, the results of our additional test on the