량발생액은 회계분야에서 기업의 이익조정 현상을 분석하는 측정치로 자주 사용된다. 또한 회계투명성 또는 신뢰성, 감사품질의 측정치로 사용되기도 한다. 그런데, 국내 재량발생액 연구에서 통제변수는 일관된 결과를 보이지 못하는 경우가 발생하고, 연구자 또한 통제변수는 관심변수가 아니므로 구체적 언급을 하지 않는 경우도 있다. 본 연구는 국내 재량발생액 연구에서 자주 사용되는 통제변수에 초점을 맞추어 분석하였다. 먼저, 문헌검토를 통해 국내 재량발생액 연구에서 통제변수가 어떠한 결과를 보였는지를 분석하였다. 58편의 선행연구에 대한 문헌검토 결과, 손실여부, 전기발생액, 기업규모, 부채비율, 대주주지분율, 외국인지분율, 성장기회, 대형감사법인여부가 통제변수로 주로 사용되고 있었다. 재량발생액에 음(-)의 영향을 미칠 것으로 예상되는 통제변수(손실여부, 전기발생액, 외국인지분율, 대형감사법인여부) 중 손실여부는 50%, 전기발생액은 30.5%, 외국인지분율은 8.2%, 대형감사법인여부는 14%가 유의한 음(-)의 계수 값을 보였다. 재량발생액에 양(+)의 영향을 미칠 것으로 예상되는 통제변수(부채비율, 성장기회) 중 부채비율은 17.7%, 성장기회는 43.1%가 유의한 양(+)의 계수 값을 보였다. 문헌검토 결과, 통제변수는 선행연구의 예상과 다른 결과를 보이는 경우가 많았는데, 본 연구는 [이익 – 영업활동으로인한현금흐름 = 발생액]의 관계에 의거, 새로운 시각에서 통제변수를 해석해 보았다. 즉, 통제변수가 발생액과 함수관계를 갖는 이익 및 영업활동으로인한현금흐름을 대리하는 변수로서 역할을 할 것으로 예상하고, 이를 분석하였다. 1995년부터 2013년까지 비금융업에 속하는 12월 결산 유가증권시장 6,061 기업-년 표본과 코스닥시장 5,437 기업-년 표본을 분석한 결과, 본 연구의 주장대로 통제변수가 이익 또는 영업활동으로인한현금흐름을 대리하여 체계적인 부호의 변화를 보이며 재량발생액에 영향을 미치는 것으로 나타났다. 이는 통제변수가 이익 또는 영업활동으로인한현금흐름을 대리하는 변수에 불과함을 의미하고, 문헌검토 결과에서 통제변수가 선행연구의 예상과 다른 결과를 보인 원인일 수 있다. 또한, 이익조정 유인에 근거해서 사용하는 선행연구의 통제변수가 이익조정 유인을 포착하고 있는가에 대한 의문을 제기할 수 있음을 의미한다.

Discretionary accruals are often used as a measure to analyze the phenomenon of earnings management in the accounting field. Also sometimes they are used as a measure of accounting transparency and audit quality. Like this, discretion accruals play an important role in the accounting studies. Nevertheless, it is difficult to find a literature research on discretionary accruals in Korea. Therefore, the first purpose of this study is to review the literature on discretionary accrual researches in Korea. Especially, this study focuses on the control variables that are frequently used in the discretion accrual studies. This study reviewed 58 domestic previous researches that designed the regression model of the discretionary accruals as dependent variable. According to the literature review, the debt ratio and firm size was the most commonly used as control variables. Previous studies mainly used loss dummy variable if net income is negative, previous accruals, firm size, debt ratio, the ration of the largest shareholder ownership, the ratio of the foreign ownership, growth opportunities and big4 dummy variable if audit firm is big4 as the control variables. According to the general theory, loss dummy variable, previous accruals, the ratio of the foreign ownership, big4 dummy variable are expected to affect discretionary accruals negatively. But, 50% of the loss dummy variable, 30.5% of the previous accruals, 8.2% of the ratio of the foreign ownership and 14% of the big4 dummy variable affected discretionary accruals negatively in the review of the literature researches of this study. Also, according to the general theory, debt ratio and growth opportunities are expected to affect discretionary accruals positively. But, 17.7% of the debt ratio and 43.1% of the growth opportunities affected discretionary accruals positively in the review of the literature researches of this study. In other words, the general theory has shown a significant difference with the empirical results. Therefore, the second purpose of this study was to find the reason of the difference between the general theory and the empirical results. This study tries to find the reason of the difference in the relationship of [earnings – cash flows from operations = total accruals]. Because discretionary accruals estimating model like modified Jones model(Dechow et al. 1995) have low R2, much portions of the total accruals remain the discretionary accruals. So, earnings and cash flows from operations can effect discretionary accruals like total accruals. Eventually, we can infer that control variables like loss dummy variables, previous accruals, firm size, debt ratio, the ration of the largest shareholder ownership, the ratio of the foreign ownership, growth opportunities and big4 dummy variable would surrogate earnings and cash flows from operations to affect discretionary accruals. For empirical tests, this study investigates if the regression model of discretionary accruals as dependent variables excludes cash flows from operations or earnings as independent variables, cash flows from operations or earnings are proxied by other control variables like loss dummy variables, previous accruals, firm size, debt ratio, the ration of the largest shareholder ownership, the ratio of the foreign ownership, growth opportunities and big4 dummy variable. Discretionary accruals were estimated using the cross-sectional modified Jones model(Dechow et al. 1995), ROA model(Kothari et al. 2005) and CFO model(Rees et al. 1996). The sample consisted of 6,061 firm-year samples in KOSPI and 5,437 firm-year samples in KOSDAQ from 1995 to 2013. This study can be summarized as follows: First, when the regression model of discretionary accruals as dependent variables excludes cash flows from operations as independent variables, cash flows from operations are proxied by control variables. Second, when regression model of discretionary accruals as dependent variables excludes earnings as independent variables, earnings are proxied by control variables. Like the suggestion of this study, control variables have shown systematic changes in regression coefficient sign to surrogate cash flows from operations or earnings. This means that it is doubtful that control variables that used on the basis of earnings management incentives are to control the earnings management incentives. The results of this study might provide the starting point of the discussion of whether it is reasonable to use the cash flows from operations or earnings as independent variables in the regression model of the discretionary accruals, whether it is reasonable to use the cash flows from operations or earnings as independent variables to estimate the discretionary accruals. In addition, the survey of the Graham et al.(2005) that CEO manages earnings by using discretionary expense like research and development expense and advertising expense can be the evidence that academic researches of the earnings management(i.e. discretionary accruals) differ from practices.

Discretionary Accruals, CFO, ROA