acc & ec
DESCRIPTION
Two tales on nonprofit earnings management: - accounting and (micro)economics - empirical analysis Marc Jegers. Acc & Ec. Focus on agency gap between board and management (Steinberg 1986, Rand J. of Ec.). Acc & Ec. Manipulations: financial statement: - real: b R accounting: b A - PowerPoint PPT PresentationTRANSCRIPT
Two tales on nonprofit earnings management:
- accounting and (micro)economics- empirical analysis
Marc Jegers
Pag.2Navorsingsdag ES 2009
Acc & Ec
Focus on agency gap between board and management
(Steinberg 1986, Rand J. of Ec.)
Pag.3Navorsingsdag ES 2009
Acc & Ec
Manipulations:– financial statement:
- real: bR
- accounting: bA
- cost accounting:- profit vs. nonprofit: Δip- program vs. adm.+fundr.: Δis
Agency: k (0: budget max.; 1: service max.)
Pag.4Navorsingsdag ES 2009
Acc & Ec: real vs. accounting
earnings disclosed:
ΠM = Π + bA + bR – ½ αRbA² - ½ βRbR²
utility function:
S + F - ½ αPbA² - ½ βPbR² - k(½ αRbA² + ½ βRbR²)
Proposition 1: If subsidies and gifts are affected by the overall earnings disclosed by a nonprofit organization, a larger principal-agent gap will result in more important earnings manipulations.
Pag.5Navorsingsdag ES 2009
Acc & Ec: taxable vs. non-taxable
earnings disclosed:
ΠM = Π + tΔipCI – γR| Δip|
utility function:
F + S + t ΔipCI – kγR|Δip| - ½ γP(Δip)²
Proposition 2: A deeper principal-agent gap results on average in more manipulations as to the allocation of indirect costs to taxable activities.
Acc & Ec
Pag.6Navorsingsdag ES 2009
Acc & Ec: service vs. non-service
earnings disclosed:
ΠM = Π – δR|Δis|
utility function: F + S - kδR|Δis| – ½ δP(Δis)²
Proposition 3: A deeper principal-agent gap results on average in more manipulations as to the allocation of the mission related indirect costs to program activities.
Acc & Ec
Pag.7Navorsingsdag ES 2009
Acc & Ec: risk aversion
after having introduced risk aversion, maximising certainty equivalent utility function:
Proposition 4: Both risk-neutral and risk-averse managers of nonprofit organizations will, on average, manipulate more the profits reported, the costs allocated to profit activities, and the costs allocated to program activities, the wider the gap between their objectives and the board’s objectives.
Pag.8Navorsingsdag ES 2009
Empirical analysis
NOT direct test of previous hypotheses
Hypothesis 1 (zero profit hypothesis): Nonprofit organisations prefer to manage earnings in order to disclose zero profits.
Hypothesis 2: Under the zero profit hypothesis, unmanipulated earnings and manipulated earnings are negatively correlated.
Hypothesis 3: Nonprofit organisations receiving relatively more subsidies, will be more inclined to manipulate earnings than nonprofit organisations receiving relatively less subsidies.
Acc & Ec
Pag.9Navorsingsdag ES 2009
Empirical analysis
Hypothesis 4: Nonprofit organisations characterised by a relatively deeper agency gap, will be more inclined to manipulate earnings than nonprofit organisations characterised by smaller agency gaps.
Hypothesis 5: There is a negative relationship between nonprofit organisations’ indebtedness and the level of earnings manipulations.
Hypothesis 5’: There is a positive relationship between nonprofit organisations’ indebtedness and the level of earnings manipulations.
Pag.10Navorsingsdag ES 2009
Empirical analysis
Sample: 1,054 financial statements (2007)
Method: probit (yes/no), heteroscedastic-consistent OLS (amount)
Pag.11Navorsingsdag ES 2009
Empirical analysis
Dependent: accruals (exc., fin.+exc., all)
Independent: earnings before ‘manipulation’, relative subsidies, agency gap, financial debt/TA, ln Assets, industry
Pag.12Navorsingsdag ES 2009
Empirical analysis: before
-1e6 -800000 -600000 -400000 -200000 0 200000 400000 600000 800000 1e6
5
10
15
20
25
frequency
EBM
Pag.13Navorsingsdag ES 2009
Empirical analysis: after
-1e6 -800000 -600000 -400000 -200000 0 200000 400000 600000 800000 1e6
5
10
15
20
25
30
35
40
45 frequency
EBEF
Pag.14Navorsingsdag ES 2009
Empirical analysis
Table 3: Correlations between absolute values of manipulated and unmanipulated earnings
Variables correlated Pearson correlation (%) Observations absEBM-absME 63*** 844 absEBM-absME (≠ 0) 64*** 665 absEBEF-absMEFE 45*** 844 absEBEF-absMEFE (≠ 0) 62*** 250 absEBE-absMEE 43*** 844 absEBE-absMEE (≠ 0) 70*** 166
Note: *,**,*** : 10%, 5%, 1% significance levels respectively (one-sided)
Pag.15Navorsingsdag ES 2009
Empirical analysis: one of the tables
Table 6: Absolute levels of manipulated earnings (Panel A: overall manipulated earnings (absME): total sample (n=844), restricted sample (n=665))
(14) (15) (16) (17) (18) (19) (20) (21) Constant (1,000) 1,081** 1,469** -3,169*** 1 4,320*** -17 -1,302*** -1,791*** RELSUB 5,286 2,917 13,049 5,110 15,699 2,427 AG 21,406** 31,553** 52,196*** -10,419 69,391*** -12,052 DEBT (1,000) 102 110 41 114 66 122 absEBM 0.15** 0.16** 0.17** 0.18** int:RELSUBabsEBM -0.09 -0.07 int:AGEabsBM 0.22*** 0.20** int:DEBTabsEBM 0.63** 0.62** absrEBM (1,000) 311 846** SIZE (1,000) 160** 212** 477*** 634*** 196*** 270*** industry dummies yes yes yes yes yes yes yes yes N 844 665 844 844 665 665 844 665 R² 46 53 22 41 26 43 45 46 Adj. R² 42 44 21 40 25 42 43 44
Pag.16Navorsingsdag ES 2009
Empirical analysis
Conclusions:
Effect of size, agency gap, debt.