http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
- 中文 을 입력하시려면 zhongwen을 입력하시고 space를누르시면됩니다.
- 北京 을 입력하시려면 beijing을 입력하시고 space를 누르시면 됩니다.
RISS 인기검색어
- 검색키워드
- 저자 : 안치현 ( Chihyoun Ahn ) (검색결과 1 건)
- 무료
- 기관 내 무료
- 유료
-
이익조정 측정모형으로서 발생액모형과 수익모형의 적합성, 검정력 비교연구
안치현 ( Chihyoun Ahn ),김미옥 ( Mi-ok Kim ),원현희 ( Hyun-hee Won ),정형록 ( Hyung-rok Jung ) 한국회계학회 2016 회계저널 Vol.25 No.1
본 연구의 목적은 Stubben(2010)의 수익모형이 우리나라 기업의 이익조정을 식별해 내는지 분석하는 것이다. 구체적으로 국내 상장기업의 수명주기별로 수익모형과 발생액모형의 모형적합성과 모형검정력을 비교하여 수익모형이 경영자의 이익조정을 식별해 낼수 있는 있는지를 확인하였다. 분석결과는 다음과 같다. 첫째, 회귀분석결과 4사분기 수익변화의 계수가 1~3사분기의 수익변화의 계수보다 10배 이상 컸다. 또한 분기수익모형이 연간수익모형보다 설명력이 더 높은 것으로 나타났다. 둘째, 시뮬레이션 결과 평균편의는 모든 모형에서 0에 근사한 값을 보였다. 표준편차는 수익모형들이 발생액모형들보다 작았다. 모형적합성 분석에 있어서 표본에 따라 다른 결과를 보였다. 모형의 검정력을 분석한 결과 수익조작은 수익모형이, 비용조작은 발생액모형의 검정력이 높은 것으로 나타났으며, 수익-비용조작에 있어서는 두 모형들이 비슷한 검정력을 보였다. 본 연구는 기존의 발생액모형과 수익모형의 시뮬레이션을 통해 기업의 수명주기에 따른 모형의 적합성과 검정력을 비교함으로써 수익모형의 유용성을 파악하는데 밑거름을 다졌다. 따라서 향후 이익조정 연구에서 발생액모형의 보완적 모형으로 수익모형을 소개하였다는데 시사점이 있다. This research studies whether Stubben(2010)’s revenue models identify earnings management in Korea firms. Prior studies related with earnings management generally use aggregate accrual models to detect earnings management, such as Jones model, modified Jones model, and performance matched modified Jones model. However, there have been controversial issues about discretional accrual as a proxy of earnings management. Accordingly, this study examines the usefulness of discretionary revenue models to detect earnings management in Korean firms. Stubben(2010) asserts that revenue, as the most important component of earnings, is a variable affected by the discretion of managers. This study specifically tests whether revenue models can detect earnings management of managers by comparing the specifications and powers between revenue models and aggregate accrual models for listed firms in Korea. As our contributions to prior studies, this study suggests that the specification and power of revenue models and aggregate accrual models should be investigated by corporate life cycle, and that either or both models should be complementarily used according to the characteristics of corporate cycle stages. Based on the previous studies(Anthony and Ramesh 1992; Koh and Kim 2012) suggesting that management environment, organizational structure, management strategy, and decision-making structure are the function of corporate life cycle, which also leads to a difference in the characteristics of accounting information. The results of our study are as follows: First, from results of regression analysis, in the revenue models, the coefficient on change in fourth-quarter revenues is over ten times higher than that of the change in revenues of the first three quarters. In addition, the adjusted R2 of quarter revenue model is higher than annual revenue model. Our results are similar to those of Stubben(2010) because this study considers the likelihood that revenues of the fourth quarter in the quarter revenue model are not cashed out and remain as accounts receivables at year-end. The Fama-Macbeth regression analysis in the extended annual revenue model reveals that most of variables are not statistically significant. This is inconsistent with Stubben(2010), where the coefficients of all variables in the extended annual revenue model are significant, except GRM_S Qi,t. It is considered that this inconsistency is due to the difference in the number of samples between Stubben(2010) and this study, and that multicollinearity among independent variables, including revenues, size, age and gross margin, affects the result. For this reason, caution needs to be taken in interpreting the extended annual revenue model for Korean firms. Second, from results of simulation, the mean bias of each model is approximate zero. This is consistent with the reasoning behind the model design, where the mean bias of each model needs to have an approximate zero because the mean of discretionary accruals(or revenues) is zero. The standard deviation of the revenue models is lower than that of the aggregate accrual models. In revenue models, the standard deviation of the revenue model is lower than that of the extended annual revenue model on average. Since a model with a lower standard deviation is more likely to detect manipulation, it is anticipated that revenue models have a higher possibility to detect manipulation than aggregate accrual models. In the specification of model analyses, specification is based on models without under- or over-reject. Revenue models are advantageous in growth firms, mature stage and stagnant stage, while aggregate accrual models are advantageous in growth stage and mature/stagnant stage. In the entire firms and growth/mature stage, both revenue models and aggregate accrual models are well specified. In addition, the power of models is analyzed based on the highest detection rate. The results show that the power of revenue models is higher for revenue manipulation, while the power of aggregate accrual models is higher for expense manipulation, and that both models have a similar power in revenue-expense manipulation. Taken together, it is found that firms with more potential to grow are more interested in increasing revenue, while firms in the mature stage are more interested in saving costs and increasing profits. This indicates that revenue models can be used as complementary models of aggregate accrual models. This study lays the groundwork for identifying the usefulness of revenue models by comparing the specification and power through a simulation of aggregate accrual models and revenue models. Therefore, this study has an implication that it introduces revenue models as the complementary models of aggregate accrual models. It is expected that the study would be a starting point for the further studies on revenue models. The limitations of this study are as follows. First, the considerably small number of samples in this study with 5,167 firm-years, compared to 70,580 firm-years in Stubben(2010), is likely to affect the statistical significance of the results. Thus, a future analysis with larger samples will enhance its comparability with Stubben(2010)’s. Second, this study can not develop a model that considers quarterly characteristics and firm characteristics. The development of a model that considers the characteristics variables with seasonality will increase the explanatory power of revenue models. Third, Stubben(2010) regards firms with accounting fraud investigated by SEC as proxies of accounting manipulation firms for analysis purpose. This study, however, can not conduct the same analysis with Stubben(2010) because the number of samples in firms subject to enforcement actions, which are proxies of actual accounting manipulation firms in Korea, is not large enough.
연관 검색어 추천
이 검색어로 많이 본 자료
활용도 높은 자료
