순이익 수치의 비정상 분포를 통한 이익관리의 고찰|RISS 상세보기

다국어 초록 (Multilingual Abstract)

Earnings management is exercised with various intentions(Healy, 1985; Trueman and Titman, 1988; Jones,1991; Scholes et al.,1992; Defond and Jiambalvo, 1994). Typically, prior studies have showed that earnings are managed in order to meet certain targets or thresholds. Those targets or thresholds are aimed at: (1) reporting positive earnings (i.e. loss aversion); (2) sustaining recent performances (i.e. report a higher earnings figure than previous year); (3) meeting investors` expectations. Many of the prior studies analyzed the effects on earnings management, with no regards to the distribution of the earnings numbers. Carslaw(1988) and Thomas(1989) expanded their research on earnings management to the extent of the distribution of earnings figures. They suggested that managers tend to round-up the first left digit numbers of the reported earnings (i.e. for companies reporting profits). According to Carslaw(1988), this kind of behavior is inspired by the so-called ``1.99`` phenomenon. That is, even though the difference is only $0.01, the consumer perceives the price of $1.99 to be considerably cheaper than $2.00. In this study, we tried to study whether managers of the Korean listed companies also engage in this kind of ``earnings rounding-up behavior``. Analogous to the earlier studies, our study compares the observed with the expected frequencies of the second left digit in the reported earnings. As the first left digit numbers would be more easily remembered and considered of greater significance by the information users, the managers want to emphasize the first left digit earning numbers. So the managers who intend to increase the reported earnings would round up from the second left digit earnings numbers(SDEN) to the first left digit numbers. The paper is intended to test the following two hypotheses. The first is whether SDEN is equally observable or not in both positive or negative earnings circumstances. The second hypothesis is whether the anomalies of the SDEN are related to earnings management or not. The test sample consists of 2,951 firm-year data, selected from the firms listed on the Korea Stock Exchange (KSE) of 1995 ～ 2003. To test the first hypothesis, the paper estimates the expected distribution of SDEN by using Benford`s law(1938) which offers the expected probabilities. In our test, the observed distribution of SDEN of the positive income numbers was different from that of the negative income numbers. That is, profit reporting firms show a lower SDEN(0, 1, and 2) relative to the loss reporting firms. And higher numbers (7, 8 and 9) are more frequently observable in loss reporting firms` SDEN than in profit reporting firms. Our evidence is consistent with the results of Thomas (1989), which observed more zeroes and fewer nines than expectation in the second left digit in profit numbers. But in loss numbers, more nines and fewer zeroes are observed. The findings suggest that firms exercise discretion to increase positive earning numbers and decrease loss numbers by slight rounding up the second left digit numbers. However, there has not been distinct evidence that the documented unusual pattern means earnings management. We extend Carslaw (1988) and Thomas (1989) to the earnings management study by using the managers` discretionary accruals (DA). To test the second hypothesis, the paper estimates DA by the version of modified Jones model (1995) as the proxy of earnings management. This model implicitly assumes that all changes in credit sales and PPE (property plant and equipment) result from earnings management. If lower SDEN(0, 1 and 2) has a significantly positive correlation to DA, managers would round up ex-ante higher SDEN in order to increase reported income numbers. At all samples (2,951), the coefficient of SDEN (1 if SDEN are 0, 1 and 2; 0 otherwise) is positively significant. Our results suggest that managers of the Korean listed companies tend to round-up the reported earnings, in a way that increases the first left digit numbers by rounding up the second left digit numbers. The results in our paper are consistent with the prior studies. The major contribution of this study, however, is that it introduces discretionary accruals in this line of research. Our results clearly suggest that discretionary accruals are used in order to round-up reported earnings figures. Moreover, discretionary accruals enable us to increase the power and efficiency of the tests used in previous studies.

번역하기

참고문헌 (Reference)

1 Thomas, "Unusual Patterns in Reported Earnings" 773-787, 1989

2 Benford, "The law of anomalous numbers Proceedings of the American Philosophical Society 78" 1938

3 Park, Jong-sung, "The Reform of Accounting Systems and the Usefulness of Accounting Information" Printed in Korean 28 (28): 105-134, 2003

4 Subramanyam, B, "The Pricing of Discretionary Accruals" 22 : 249-281, 1996August-December

5 Healy, "The Effect of Bonus Schemes on Accounting Decisions Journal of Accounting and Economics 7" 85-107, 1985

6 Nigrini, M, "Taxpayer Compliance Application of Benford′s Law" 18 (18): 92-91, 1996

7 Wiley, "Statistical method for rates and propositions" 1981

8 Das, S, "Rounding-up in reported EPS, behavioral thresholds, and earnings management" 35 : 31-50, 2003

9 Gaber, "Prices as Indicator of Quality Report on an inquiry" eco (eco): 43-70, 1966

10 Ley, E, "On the peculiar distribution of the U.S. stock indices digit" 50 : 311-313, 1996