Overall, the results suggest that real earnings management is associated with impaired earnings quality through abnormal reductions in discretionary costs. Keywords: real profit management; an abnormal reduction in discretionary spending; cash flow manipulation; research and development; advertising costs; persistence of earnings; quality of earnings;. Second, the abnormal reduction in discretionary costs provides a cleaner setting for investigating the effect of real earnings management on earnings quality.
Since real earnings management, through the abnormal reduction of discretionary spending, reduces earnings persistence, it can affect the relationship between current earnings and future cash flows. H1: The degree of management of real earnings through the abnormal reduction in discretionary spending is negatively associated with earnings persistence. If real earnings management reduces earnings persistence through the abnormal reduction in discretionary spending, this should reduce the informativeness of current earnings about future cash flows.
Research design
H3: Real earnings management through the abnormal reduction in discretionary spending weakens the relationship between current earnings and future cash flows. 2008) document that the level of real earnings management activities increased after the transition to SOX because real earnings management techniques are likely to be more difficult to detect. After the transition to SOX, accrual-based earnings management therefore becomes more expensive, and therefore companies can switch to real earnings management. Since real earnings management is more expensive, the shift to real earnings management after SOX increases the average cost of earnings management and consequently has a more pronounced effect on earnings quality.
H4a: The effect of real earnings management through abnormal discretionary cost reduction on earnings persistence is stronger in the post-SOX period than in the pre-SOX period. H4b: The effect of real earnings management through abnormal discretionary expense reduction on the link between current earnings and future cash flows is stronger in the post-SOX period than in the pre-SOX period. One of the difficulties in identifying real earnings management is distinguishing between observed operating activities (such as reducing R&D expenditures) and attempts to increase current firm earnings from such activities as optimal firm choices.
Conceptually, if managers engage in real earnings management by cutting one or more types of discretionary spending, these firms will exhibit abnormally low discretionary spending. For each firm year, the abnormal level of discretionary spending is the actual discretionary spending minus the forecast values of discretionary spending. Abnormal R&D expenditure is the difference between actual R&D expenditure and predicted R&D expenditure using coefficients estimated from equation (2) and then multiplied by (-1) so that a higher abnormal level of R&D expenditure implies more real earnings management.
13 I multiply (-1) by the abnormal level of discretionary spending to facilitate the interpretation of the results so that a higher abnormal level of discretionary spending means more real earnings management. The means and medians of the individual real earnings management proxies are close to zero, consistent with previous literature (Cheng et al., 2016).
Empirical analysis
16 and examine whether the impact of actual earnings management on earnings persistence affects the relationship between current earnings and future cash flows. In Section 4.4, I investigate whether the SOX transition affects the effect of actual earnings management on earnings persistence and its ability to predict future cash flows. A significantly negative α3 is expected to indicate that actual earnings management is negatively related to earnings persistence.
The results show that managing real earnings by abnormally reducing discretionary costs significantly reduces earnings persistence. H2 predicts that real earnings management by abnormally reducing discretionary costs affects the persistence of operating cash flows more than accruals. In model (4) of Table 4, where RM is equal to management measures with actual earnings achieved by reducing PSA-costs (RM_SGAi,t), the coefficients of the interaction term RMi,t*ACCi,t and RMi,t*CFOi,t both significantly negative, which suggests that the management of actual profit through PSA-costs affects both accruals and cash flows from operations.
20 the third half predicts that real earnings management through abnormal reduction in discretionary spending weakens the link between current earnings and future cash flows. Overall, the results show that real earnings management through abnormal reduction in discretionary spending weakens the link between current earnings and future cash flows. In summary, the results suggest that in post-SOX periods, the effect of managing real earnings through R&D, SG&A, and total discretionary spending reduces to a greater extent the persistence of current earnings and its informativeness of future cash flows.
I then estimate equation (8a) and (8b) to test whether the negative effect of real earnings management through abnormal reduction in discretionary spending is stronger for firms that meet or simply exceed earnings targets. Real earnings management affects both accruals and cash flows, and it is possible that the effect of the abnormal reduction in discretionary spending on earnings persistence and its informativeness about future cash flows is driven by the effect of earnings management accruals (approximated by large accruals). In my main tests, I follow the previous literature to measure the extent of real earnings management through abnormal reduction in discretionary spending.
However, my findings suggest that managing real earnings by abnormally reducing discretionary spending reduces earnings persistence on average.
Conclusions
In my main tests, I use different samples for the tests of each type of discretionary spending to maximize my sample size.28 In sensitivity tests, I construct a common sample that has non-missing data for all types of discretionary spending to re-run all my tests. The purpose of these tests is to check whether my results are sensitive to different samples. Examining a sample over a period of 41 years, I find that earnings persistence falls short in all my measures of real earnings management.
The results are consistent with prior research documenting a negative relationship between actual earnings management and future corporate performance. In testing whether this effect on earnings persistence is due to cash flows or accruals, I found that managing actual earnings with abnormal reductions in all types of discretionary spending has a greater effect on cash from operations than accruals. Furthermore, my results show that less persistent earnings as a result of actual earnings management are less informative about future cash flows.
This study contributes to the literature on the investigation of earnings quality under a certain type of earnings management. Prior study documents a trend for managers to increase the use of actual operational activities to manage earnings following the passage of SOX. The results of my study indicate that real earnings management negatively affects earnings quality, especially the quality of cash flow.
Existing research examining earnings quality focuses largely on accrual quality, implicitly assuming that cash flows are free from manipulation. My findings suggest that future researchers should consider both accruals and real earnings when conducting earnings quality research.
STDEBTi,t = change in debt in short-term liabilities between year t – 1 and year t, Depreciationi,t = depreciation in year t. ADVi,t Advertising Expenditure ASSETSi,t Total Assets. CFOi,t Cash flows from operations, equal to Ei,t – ACCi,t. DIVDUMi,t It is equal to 1 when the company pays dividends in year t. Ei,t Net profit before extraordinary items in year t scaled by average assets LOSSi,t It is equal to 1 if the company has a loss in year t. OANCFi,t Cash flows from operations taken directly from the cash flow statement.
ASSETSi,t are total assets; Ei,t is net profit before extraordinary items; OANCFi,t are the operating cash flows from the cash flow statement; LOSSi,t is equal to 1 if companies have a loss in year t, and zero otherwise; RM_R&Di,t is the abnormal level of expenditure on research and development; RM_ADVi,t is the abnormal level of advertising expenditure; RM_SGAi,t is the abnormal level of PSA-costs; and RM_TDISXi,t is the abnormal level of total discretionary expenditure. DIVDUMi,t is equal to 1 when company i pays dividends in year t, and 0 otherwise; LOSSi,t equals 1 if company i has a loss in year t, and 0 otherwise; ACCi,t are total accruals. The effect of real earnings management on the link between current earnings and future cash flows.
OANCF is cash flow from operations taken directly from the cash flow statement; Ei,t is net profit before extraordinary items. Log(ASSETS)i,t is the natural log of total assets; DIVi,t is dividend, DIVDUMi,t is equal to 1 when company i pays a dividend in year t, and 0 otherwise; LOSSi,t equals 1 when company i has a loss in year t, and 0 otherwise; ACCi,t are total accruals. DIVDUMi,t is equal to 1 when company i pays dividends in year t, and 0 otherwise; LOSSi,t equals 1 when company i has a loss in year t, and 0 otherwise; ACCi,t are total accruals; SOXi,t is equal to 1 for the period 2002 onwards; 0 otherwise.
OANCFi,t are cash flows from operations taken directly from the cash flow statement; Ei,t is net profit before extraordinary items. Log(ASSETS)i,t is the natural log of total assets; DIVi,t are dividends, DIVDUMi,t is equal to 1 when company i pays dividends in year t, and 0 otherwise; LOSSi,t equals 1 if company i has a loss in year t, and 0 otherwise; ACCi,t are total accruals;. DIVDUMi,t is equal to 1 when company i pays dividends in year t, and 0 otherwise; LOSSi,t equals 1 if company i has a loss in year t, and 0 otherwise; ACCi,t are total accruals; MBTi,t is equal to 1 if firms only reach the target of zero profit or last year's earnings or if current earnings are greater than zero earnings or last year's earnings increased by total assets by 0.01, and 0 otherwise.
DIVDUMi,t is equal to 1 when the firm pays dividends in year t, and 0 otherwise; LOSS,t is equal to 1 when firm i has a loss in year t., and 0 for others; ACCi,t is total accrual; DAi,t is equal to 1 for firms with discretionary accruals in the top quintile of absolute residual values estimated by the modified Jones model, and 0 otherwise.
