Earnings Management Strategies: To Conform or Not to Conform

1
Earnings Management Strategies To Conform or Not
to Conform?

• Brad Badertscher Notre Dame University
• John Phillips University of Connecticut
• Morton Pincus University of California, Irvine
• Sonja Olhoft Rego University of Iowa

2
Research Questions

• We investigate the extent to which pretax
  earnings management has current income tax
  consequences.
• Do firms rely on more book-tax conforming or
  nonconforming earnings management strategies?
• Which firm-specific characteristics impact the
  choice between conforming and nonconforming
  earnings management strategies?

3
Motivation

• Our paper is motivated by Healy and Wahlens
  (1999) call for a better understanding of how
  firms manage earnings upward, and by two streams
  of research
• Studies that investigate earnings management
  activities that have current tax consequences.
• E.g., Erickson, Hanlon, and Maydew 2004.
• Studies that investigate earnings management that
  does NOT have current tax consequences.
• E.g., Phillips, Pincus, and Rego 2003.

4
Methodology

• We utilize a sample of firms that restated
  earnings due to accounting irregularities between
  1997 and 2002.
• Misstatement firms
• We use originally reported and restated financial
  statement data to measure the amount of total
  earnings management, as well as the amounts of
  conforming and nonconforming earnings management.
• Misstatement firms that managed earnings upward
  in a book-tax
• Conforming manner should have restated their
  current tax expense downward (along with their
  pretax income).
• Nonconforming manner should have restated their
  deferred tax expense downward (along with their
  pretax income).

5
Research Findings

• We first consider the prevalence of conforming
  and nonconforming earnings management by
  misstatement firms.
• We find that nonconforming EM dominates our
  sample of misstatement firms.
• Problematic finding since book-tax differences
  likely increase the probability of detection, so
  our empirical tests suffer from a self-selection
  bias.

6
Research Findings

• We then investigate which firm-specific
  characteristics impact the choice between
  conforming and nonconforming EM strategies.
• We propose that firms tradeoff
• the expected benefits (i.e., lower current tax
  costs)
• against the expected costs (i.e., higher
  probability of detection) associated with
  nonconforming EM strategies.
• We find that firms with
• NOL carryforwards,
• a Big 4/5/6 auditor,
• or that engaged in accounting fraud,
• rely on relatively more book-tax conforming
  earnings management strategies.

7
Contributions

• We examine the choices managers make when
  deciding how to manage earnings upward.
• These choices systematically reflect differences
  in the costs and benefits of using book-tax
  conforming vs. nonconforming accruals.
• Managers are more likely to engage in book-tax
  conforming EM when such EM does not generate
  current tax costs (i.e., NOL c/fs) and to reduce
  the probability of detection.
• This paper also provides a comprehensive
  framework for understanding how income tax
  accounts reveal different types of pretax
  earnings management.

8
Background

• We build on Erickson, Hanlon, and Maydew, 2004.
• How much tax are firms willing to pay to manage
  earnings upward?
• Less than 35 on the dollar!
• Examples of conforming earnings management
• Accelerating revenue recognition (e.g., channel
  stuffing),
• Manipulating year-end cut-offs (e.g., LIFO
  inventory purchases and revenue recognition), and
• Examples of nonconforming earnings management
• Reducing reserve accounts, (e.g., allowance for
  doubtful accounts, restructuring reserves,
  contingent liabilities),
• Extending the useful lives of depreciable assets,
  and
• Aggressively recognizing revenue from unearned
  revenue.

9
GAO Restatement Sample

• We rely on the 2002 GAO report to obtain a sample
  of firms that likely managed earnings upward.
• This report only includes restatements due to
  accounting irregularities that caused material
  misstatements of financial results.
• The accounting irregularities in the GAO report
  reflect
• aggressive accounting practices,
• intentional and unintentional misuse of facts
  applied to financial statements,
• oversight or misinterpretation of accounting
  rules,
• and fraud.
• The restatement data allow us to hold constant
  the impact of tax planning (and measurement
  error) on current and deferred tax expense.

10
Tax Benefit / Detection Cost Tradeoff

• Our paper contains a simple model of the
  tradeoffs between conforming and nonconforming EM
  strategies.
• We predict that firms should engage in relatively
  more book-tax conforming EM when
• the expected benefits of nonconforming EM are
  less than the expected costs of such earnings
  management.
• So we need proxies for the expects costs and
  benefits of nonconforming EM!

11
Tax Benefit / Detection Cost Tradeoff

• Proxies for the expected benefits associated with
  nonconforming EM.
• NOL carryforwards.
• High free cash flow.
• Proxies for the expected costs associated with
  nonconforming EM.
• Big 4/5/6 auditor.
• Analyst following.
• Fraudulent activity.
• Proxied by fraud and/or managerial turnover.

12
Earnings Management Proxies

• We first compute differences between originally
  reported and restated financial data (scaled by
  total assets)
• PTI_EM, CTE_EM, DTE_EM, DNOL_EM, and DVAA_EM.
• We then adjust CTE_EM and DTE_EM for measurement
  error caused by changes in NOL carryforwards and
  the VAA
• CONFORM_EM (CTE_EM - ?NOL_EM) / STR
• NONCONFORM_EM (DTE_EM ?NOL_EM - ?VAA_EM) /
  STR
• Our latest empirical tests rely on federal and
  foreign data to remove the impact of state taxes
  from our proxies.
• We compute the proportions of conforming (C_RATE)
  and nonconforming (NC_RATE) earnings management
  by scaling the above measures by the total amount
  of EM.

13
Earnings Management Proxies

• In limited circumstances, firms can manage their
  earnings in a manner that generates a permanent
  book-tax difference.
• E.g., Managing earnings in a foreign subsidiary.
• The permanently nonconforming portion of total EM
  is a plug number
• PERM_NC_EM TOT_EM CONFORM_EM NONCONFORM_EM
• Thus, we also run our tests where NC_RATE is the
  total amount of nonconforming EM, rather than
  just the temporarily nonconforming portion.

14
Choice between Conforming and Nonconforming
Earnings Management Strategies

• NC_RATE b0 b1NOL b2HI_CASH b3BIG_AUDIT
  b4AF b5FRAUD b6INV_MILLS CONTROLS e
• Heckman two-step self-selection model
• RESTATE b0 b1DTE b2MJAC (OTHER PREDICTORS
  OF RESTATEMENTS) e
• We extract the inverse mills ratio and include it
  in the NC_RATE regression.

15
TABLE 2 RevisedDescriptive Statistics for
Misstatement Firms

• Revised table relies on federal and foreign tax
  components to estimate rates of conforming and
  nonconforming EM, rather than total current or
  deferred tax expense.
• Reduces measurement error associated with state
  and foreign taxes.

16
TABLE 3 RevisedC_RATE and NC_RATE by Reason
for Misstatement
17
TABLE 3 RevisedC_RATE and NC_RATE by Sample
Partitionings
18
TABLE 5 Revised
19
Concluding Remarks

• Nonconforming EM dominates our sample of
  misstatement firms.
• However, NOL carryforwards, use of a Big 4/5/6
  auditor, and the presence of fraud caused our
  sample of misstatement firms to engage in
  relatively more book-tax conforming EM.
• Thus, the rate of current taxes paid in Erickson,
  Hanlon, and Maydew 2004 may not generalize to a
  larger sample of firms.
• In general, our results are consistent with firms
  trading off the expected benefits of
  nonconforming EM strategies against the expected
  costs of such strategies.