기업들은 가능하다면 적자보다는 흑자를 보고하고 싶어 할 것이다. 왜냐하면 적자를 보고하는 기업들에 대해서는 이해관계자들이 별로 호의적이지 않아 기업은 이러한 이해관계자들과의 거래에서 높은 비용을 지급해야 할 가능성이 있기 때문이다. 이에 따라 적자가 발생한 기업에서 경영자들은 적자를 흑자로 조정하여 보고하려는 동기를 가지게 될 것이다. 중간보고의 경우, 연차보고에 비해서 적자회피 이익조정유인이 다를 수 있는데, 이는 중간보고에서는 이익을 조정하려는 유인과 이에 수반되는 비용이 연차보고서의 경우와 다르기 때문이다. 예를 들면, 중간보고이익에 대한 규제와 감시가 연차보고의 경우보다 덜 엄격하기 때문에 연차보다 중간보고기간의 이익조정을 보다 용이하게 수행할 가능성이 있다. 또한 이러한 유인과 비용은 분기별로도 상이할 가능성이 있다. 본 연구는 크게 두 가지 목적을 달성하기 위해 수행되었다. 첫째로 분기별로 적자회피 이익조정 현상이 어떻게 다르게 나타나는지를 검증하고자 한다. 둘째로 4분기의 이익보고 행태가 개별1~3분기 이익조정의 영향을 받은 역전현상의 결과로 나타난 것인지를 분석하고자 한다. 본 연구의 실증분석 결과에 의하면, 개별4분기의 적자회피 이익조정은 동일 회계기간의 이전 개별분기들에 비해서 매우 낮은 것으로 나타났다. 특히, 이러한 개별4분기의 현상은 동일회계기간의 이전분기 이익조정을 청산한 결과로 나타났을 가능성을 밝히고 있다. 이는 개별4분기를 별도로 작성 · 공시하지 않아도 되는 현행 기준서의 규정 때문일 가능성이 있다. 이러한 결과는 4분기에는 연차이익에 대한 기대치를 맞추려고 4분기 이익을 미래로 이연시키거나 그 반대로 미래이익을 4분기에서 차용하고 있다고 보고하고 있는 Das et al.(2007)의 결과와는 다르다. 한편, 1분기의 적자회피 이익조정은 다른 분기에 비해 상대적으로 빈번히 행해지고 있었다. 본 연구의 결과는 중간분기이익의 적정성을 검토한다는 측면에서 유용한 자료로 사용될 수 있을 것이며, 궁극적으로는 중간분기와 관련된 회계제도를 발전시켜 나갈 방향에 대한 자료로써도 활용될 수 있을 것이다. 특히, 이는 개별 4분기보고를 연차보고로 대체할 수 있는 현행규정을 연차보고와 더불어 개별4분기를 동시에 공시하도록 하는 규정으로 변경하는 근거를 제시하고 있다.
Companies have incentives to avoid reporting a loss because of the potentially, increased transactions costs that will result when dealing with lenders and other related parties. It is also likely that the incentive for loss avoidance in an interim period differs from the incentive for loss avoidance in the annual report because both the increased costs associated with reporting a loss and the extent of monitoring of financial reporting differ across interim and annual reports. Consequently, the extent of earnings management will differ across interim and annual reports and may also differ across the first three interim reports. This study has two objectives. First, it examines whether there is a difference in the extent of loss avoidance earnings management(LAEM) between the fourth quarter and the first three quarters. Second, it investigates whether the extent of LAEM differs across the three interim quarters. We conduct our analysis on a sample of companies listed on the Korean Stock Exchanges over the period 2003-2006. Following the Burgstahler and Dichev (1997) method for detecting LAEM, we divide the scaled(by beginning market value of equity) quarterly income distribution into intervals of 0.005 around the cutoff point “0”, and examine the frequencies in these intervals. Our results indicate a steep drop in frequency in the -0.005 interval in the first and second quarters, a smaller drop in the third quarter, and no unexpected drop in the fourth quarter (see Figure 2: Distribution of Quarterly Earnings). Our tests are based on standardized differences(SDs) between actual and expected frequencies in each interval. If there is no earnings management, actual and expected frequencies should have similar values and therefore, SDs should not significantly differ from zero. Our results also indicate that SDs are highest in the first quarter and decrease monotonically from the second to the fourth quarter. These results suggest that companies avoid reporting losses in the earlier interim quarters but not in the fourth quarter. Our results indicate that differences between the fourth quarter’s and the first, second and third quarters’ SDs are -14.69, -10.30 and -5.32 respectively, in the -0.005 interval, and are 14.36, 6.28 and 9.89 respectively, in the +0.005 interval, and each of these differences is significant at the 0.01 level. These results are consistent with our first hypothesis that LAEM in the fourth quarter is lower than LAEM in the first three interim quarters(see Table 4-4: SDs in Quarterly LAEM) Additional tests using discretionary accruals(DA) estimated from the modified Jones model provide confirmatory evidence. The difference between the fourth quarter’s and first quarter’s DA is significant at the 1% level(t=-2.70). The differences between fourth quarter DA and the second and third quarters’ DAs have t-values of -2.23 and -1.95 respectively, and these values are significant at the 1% and 10% level, respectively. These results confirm the results of the LAEM analysis and provide additional support for H1. Further analysis indicates no significant difference between first quarter and second quarter DA(t=0.76). However, the difference between first quarter and third quarter DA is statistically significant(t=1.92). This is consistent with our LAEM test results and supports our supplementary hypothesis of more frequent(greater) LAEM in the first quarter than in subsequent quarters(see Table 4-5: Differences in Quarterly Discretionary Accruals) One explanation for the low level of earnings management is that fourth quarter earnings may be affected by the reversal of earnings management in the first three quarters. However, fourth quarter earnings may also be managed to meet annual earnings targets or to avoid reporting a loss by borrowing from the forthcoming period’s earnings. We examine these two competing alternatives. Our regression results indicate that the coefficients(t-value) on DA in the first quarter of the year and the third quarter of the current year are -0.115(-3.44) and -0.058 (-1.69), and are significant at 1% and 10%, respectively. On the other hand, the coefficients for the forthcoming quarters’ DA are not significant(see Table 4-6: Regression Results of Discretionary Accruals). These results are consistent with our second hypothesis that earnings management in the fourth quarter is offset by reversals of earnings management in the first three interim quarters of the year. This finding differs from Das et al.(2007) who find that fourth quarter earnings management results from shifting earnings to or from the next period. The results of this study can be used for evaluating the appropriateness of reported quarterly earnings, and will serve as a useful input in designing accounting systems related to interim reports. In addition, this study provides evidence that supports changing the current system by requiring firms to report their fourth- quarter performance along with annual results.
