Part 4
Accruals 371 In this part, we deal with the mechanics of the empirical research on earnings management. As noted below, most of the research has identified earnings management with the detection of discretionary accruals. Hence, we first offer a short discussion of accruals. Accruals Accruals arise when there is a discrepancy between the timing of cash flows and the timing of the accounting recognition of the transaction. One notable example involves the recognition of revenues. Revenues may be recognized after customers advance cash and before total collection is assured (Table 1). Table 1 The accruals process Period 1 2 3 Event An advance from a customer Shipment of the merchandize to the customer The customer settles his account Cash flows Inflow of advance None Inflow of the final payment Accounting recognition of revenue None Recording of revenues None Accruals Increase in unearned revenues Decrease in unearned revenues and/or increase in accounts receivable Decrease in accounts receivable
372 Part 4 The advance creates a liability termed unearned revenues. The final payment decreases the asset termed accounts receivable. Over the firm s lifetime, reported revenues must equal total (gross) cash inflows, and total accruals must equal zero; that is, accrued balances of assets and liabilities reverse. In our example, the unearned revenues liability is reduced to zero upon the shipment of the merchandize, and the accounts receivable asset is reduced to zero when the customer pays his debt. The research on accruals management attempts to distinguish between accruals resulting from managed earnings and normal accruals. Definitions: Non-discretionary accruals are accruals that arise from transactions made in the current period that are normal for the firm given its performance level and business strategy, industry conventions, macro-economic events, and other economic factors. Discretionary accruals are accruals that arise from transactions made or accounting treatments chosen in order to manage earnings. 1 reversals are accruals originating from transactions made in previous periods. Note that this definition is consistent with our definition in Chap. 2 in that we wish to distinguish between normal transactions and accounting treatments and abnormal transactions and accounting treatments aimed at managing earnings, and we allow for beneficial and neutral earnings management. In some cases, discretionary accruals may be quite large. Beneish (1997), finds that in his sample of 43 firms that were subject to enforcement actions by the SEC between 1987 and 1994, the mean overestimation of earnings was 42.5% of retained earnings and 11.5% of total assets (table 2, panel B). Richardson, Tuna, and Wu (2002), whose accruals encompass both working-capital and asset acquisition accruals, find that total accruals amount to 8.7% of average total assets in their sample of 440 firm-years (of 225 firms) that restated earnings between 1971 and 2000, as compared to 3.9% of the average assets for the control group of non-restatement firms (t-statistic = 4.50). 2 1 In this definition, accrued balances are all non-cash assets and all liabilities that are not borrowings (such as bank debt, bonds, etc.). 2 Beneish s 64-firm sample contains firms with missing data. Hence, the statistics applies to 43 firms only. Beneish's sample can be criticized on the ground that it is composed of young firms, so perhaps the percentages should be attributed to size. McNichols (2000), reports that the median net income before extraordinary items for Compustat firms for the 1988 1998 period is 3.8% of the beginning of the period total assets. She observes that an average earnings management of even 1% of total assets would be material to most firms earnings.
Accruals 373 Our definition implies that the total accrued balances of firm i in period t break down as follows: EBTA it = DA it + NA it + Reversal it + EBTA i,t 1, (1a) where EBTA ik = ending accrued balances of firm i in period k, k=t,t-1; DA it = Discretionary accruals of firm i resulting from transactions and events occurring in period t; NA it = non-discretionary accruals of firm i resulting from transactions and events occurring in period t; Reversal it = reversal in period t of balances accrued by firm i in previous periods. To illustrate Eq. 1a, consider this example. Suppose that the firm makes a $200 sale. The normal credit period is 30 days. Unable to make the sale without relaxing the credit policy, the firm boosts earnings by agreeing to 50% payment in 30 days, and the remaining 50% in 60 days. The breakdown of account balances and accruals is as follows: Month 1 2 3 Total EBTA i,t-1 0 EBTA it +200 100 0 DA it + 100 +100 ND it + 100 +100 Reversal it 100 100 200 Total 0 ==== In the first month, the non-discretionary accrued balance increases by 100, and the discretionary accrued balance increases by 100. In the second month, the non-discretionary accrued balance decreases by 100, and in the following month, the discretionary accrued balance decreases by 100. The payments of 100 in months 2 and 3 are the reversals. The reversals ensure that the changes in accrued balances add up to zero. The reversal of accrued balances limits the ability to manage earnings. To illustrate, consider a case in which revenues are boosted in the current year by shipping unsolicited merchandise to customers. The merchandize is expected to be returned in the next accounting period (see Chap. 2 on Similarly, Kothari, Leone, and Wasley (2005, table 1) show that the discretionary accruals measured by the Jones model are 0.19 of total assets, as compared to average total accruals of 3.03% of total assets in their 122,798 firm-year sample collected for the 1963 1999 period. The negative value is explained by depreciation.
374 Part 4 channel stuffing ). The revenues of the next period are thus lowered by sales returns, and the accounts receivable accrued balance thus reverses. Suppose, for the purpose of the example, that in the next year unmanaged earnings and the expected reported earnings remain the same. The firm then must double the size of managed earnings to offset the negative impact of the reversal of the previous year s earnings management. In the following year, the firm must triple the managed sales, and so on. Although reversals are important, it seems that most empirical research ignores them (in Chap. 11, we detail studies that attempt to deal with reversals), largely because they are unobservable. The studies equate total accruals with the difference between the ending and the beginning accrued balances. Unfortunately, this injects a measurement error into the accruals, as in the following example. The beginning balances (BB) of cash and accounts receivable are 50 and 80, respectively. The firm conducts two sales transactions, a normal sale of $2,000, which increases non-discretionary accruals, NDA, by $200, and an aggressive channel stuffing that increase sales and discretionary accruals (DA) by $50. The beginning accrued balance of accounts receivable was paid in the current period. The change in the balances of accruals is $170 ($250 $80), which also equals the difference between earnings and cash flows ($2050 $1880), but the total of NDA and DA is $250 ($200+$50). Cash Accounts receivable Sales BB 50 BB 80 1,800 (NDA) 200 2,000 (DA) 50 50 80 80 (Reversal) EB 1,930 EB 250 2,050 EB A reversed BB of $80 created this difference. If, however, the BB was not paid in the current period, the difference between the beginning and ending balances correctly reflects the sum of NDA plus DA, as would the difference between earnings (revenue) and cash flows ($2,050 $1,800). 3 3 As we shall see, empiricists ignoring reversals implicitly correct for the omission by relating their measured accruals to the difference in sales rather than to sales. However, a measure of accruals that ignores the reversals fails to reflect all what management purposefully accrues during the current period. Consequently, such a procedure fails to calibrate the full magnitude of purposeful discretionary accruals.
The Plan of Part 4 375 The Plan of Part 4 Chapter 9 presents an analytical model of the accruals generation process in the absence of earnings management. We expand on this model and examine the effect of earnings management on the statistical properties of accruals. This framework underlies the research on both earnings management and the relationships between accruals and cash flows. Chapter 10 describes the evolution of the empirical research in accruals management up through the publication of the Jones model in 1991. The second half of Chap. 10 evaluates the efficiency of the Jones model. Chapter 11 details the progress since 1991, ending with the synthesis model of Ye (2006). We also discuss alternative methodologies: the accounting choices approach and the distributional approach. The former focuses on certain accruals that are known to have a material impact on earnings. The latter assumes that a deviation of the distribution of earnings from the normal distribution indicates earnings management. A special case is the penny approach, which tests whether firms manage the penny in earnings per share to round the number in the preferred direction.