Description
Earnings Manipulation Detection
DETECTION OF EARNINGS MANIPULATION
Purpose
?
To use financial statement analysis in the detection of earning manipulations. To assess the probability that a set of financial statements contain fraud. To Simplify the process of identifying the frauds in the financial statements.
?
?
Literature Review
?
?
?
?
It draws on advances in the accounting research literature and provides a way of integrating this knowledge. With increasing level of frauds simple Analytical procedures are not much of help. Beneish Model compares GAAP violators to aggressive accruers. Both Quantitative and Qualitative factors to be considered.
Models
?
?
Simple Analytical Procedures Sophisticated Models
Simple Analytical Procedures (APs)
?
Identifying
? ?
Unusual Relationships, and Significant Changes in financial statements
? ?
Required by SAS 56 Techniques
?
? ?
Comparison of Balance Sheets Judgmental Scanning Ratio Analysis
? ? ? ? ?
Allowance of BD/Accounts Receivable Allowance for DD/Net Sales Net Sales/Accounts Receivable Gross Margin/Net Sales Accounts Receivable/Total Assets
Sophisticated Models
?
?
?
Simultaneous inclusion of several variables Earlier models e.g. Altman (1968), Tam & Kiang (1992) Artificial Neural Network (ANN) by Green & Choi (1997)
? Lacks
explanatory capabilities
?
Beneish Probit Model
Beneish Probit Model (1997)
?
Linear Regression of variables Independent Variable = p(Earnings Manipulation) Use of cumulative normal function
?
?
Concepts
?
?
GAAP Violators: 64 Firms (1987-1993) Control Firms: 1989 Firms
? Large
discretionary accruals ? Increasing sales
?
Assessment of
? Probability
of detection of the violation through distortion of statements ? Incentive/ability to violate GAAP
1. Probability of Detection
?
Six financial statement variables:
? Days
Sales in Receivables Index ? Gross Margin Index ? Asset Quality Index ? Depreciation Index ? SG&A Index ? Total Accruals to Total Assets
2. Incentives/Ability to Violate
?
Five Financial statement variables:
? Capital
Structure ? Prior Market Performance ? Time Listed ? Sales Growth ? Prior Positive Accruals Decisions
? ?
Declining Cash Sales Dummy Positive Accruals Dummy
The Manipulation Index
?
MI =
- 2.224 + 0.221 * (Days Sales in Rec Index) + 0.102 * (Gross Margin Index) + 0.007 * (Asset Quality Index) + 0.062 * (Depreciation Index) + 0.198 * (SG&A Index) - 2.415 * (Total Accruals to Total Ast) + 0.040 * (Sales Growth Index) - 0.684 * (Abnormal Return) - 0.001 * (Time Listed) + 0.587 * (Leverage Index) + 0.421 * (Positive Accruals Dummy) - 0.413 * (Declining Cash Sales Dummy)
?
Probability of Earnings Manipulation = Normdist (MI)
Median Results
P(Manipulation)
GAAP Violators = 9.5%
Ratio Days Sales in Receivables Gross Margin Index
Control Firms = 1.1%
GAAP Violators 1.269 1.042 Control Firms 1.199 1.004
Asset Quality Index
Depreciation Index SG&A Index Total Accruals to Total Assets
0.937
0.981 0.997 0.204
0.807
1.021 0.981 0.441
Sales Growth Index
Abnormal Return Time Listed (months) Leverage Index
1.431
(0.325) 29.0 0.564
1.379
0.011 31.0 0.500
Algorithm
Beinish Probit Model
Algorithm
Algorithm
P > 9.5 Violator Firm
High Risk of Earnings manipulatio n
Probability of Earnings Manipulation
P<9.5 & P>1.1
P<1.1
Control Firm
Conclusion
90%
80%
70% 60%
Percentage of Firms
83% 76% 67%
Correctly Identified Violators (Type I Error) Incorrectly Identified Control Firms (Type II Error)
50% 40% 30% 20% 10%
45% 28.60% 20.40% 13.50% 3.60% 11.72% 5.99% 4.30% 2.94%
Cut Off Percentage
0%
Conclusion
?
Holistic Approach : Several Variables Included Inexpensive: Only 3 years of data required Simplicity
?
?
?
Including Qualitative Factors may enhance the detection power
References
?
? ?
?
?
Detecting Earnings Manipulation – A work on Benish Model. The Beneish Probit Model – Beneish1997 Simple analytical Procedures - Calderon & Green 1995 Analysis of revenue and account receivables Green & Choi www.capital9.com
doc_532379672.pptx
Earnings Manipulation Detection
DETECTION OF EARNINGS MANIPULATION
Purpose
?
To use financial statement analysis in the detection of earning manipulations. To assess the probability that a set of financial statements contain fraud. To Simplify the process of identifying the frauds in the financial statements.
?
?
Literature Review
?
?
?
?
It draws on advances in the accounting research literature and provides a way of integrating this knowledge. With increasing level of frauds simple Analytical procedures are not much of help. Beneish Model compares GAAP violators to aggressive accruers. Both Quantitative and Qualitative factors to be considered.
Models
?
?
Simple Analytical Procedures Sophisticated Models
Simple Analytical Procedures (APs)
?
Identifying
? ?
Unusual Relationships, and Significant Changes in financial statements
? ?
Required by SAS 56 Techniques
?
? ?
Comparison of Balance Sheets Judgmental Scanning Ratio Analysis
? ? ? ? ?
Allowance of BD/Accounts Receivable Allowance for DD/Net Sales Net Sales/Accounts Receivable Gross Margin/Net Sales Accounts Receivable/Total Assets
Sophisticated Models
?
?
?
Simultaneous inclusion of several variables Earlier models e.g. Altman (1968), Tam & Kiang (1992) Artificial Neural Network (ANN) by Green & Choi (1997)
? Lacks
explanatory capabilities
?
Beneish Probit Model
Beneish Probit Model (1997)
?
Linear Regression of variables Independent Variable = p(Earnings Manipulation) Use of cumulative normal function
?
?
Concepts
?
?
GAAP Violators: 64 Firms (1987-1993) Control Firms: 1989 Firms
? Large
discretionary accruals ? Increasing sales
?
Assessment of
? Probability
of detection of the violation through distortion of statements ? Incentive/ability to violate GAAP
1. Probability of Detection
?
Six financial statement variables:
? Days
Sales in Receivables Index ? Gross Margin Index ? Asset Quality Index ? Depreciation Index ? SG&A Index ? Total Accruals to Total Assets
2. Incentives/Ability to Violate
?
Five Financial statement variables:
? Capital
Structure ? Prior Market Performance ? Time Listed ? Sales Growth ? Prior Positive Accruals Decisions
? ?
Declining Cash Sales Dummy Positive Accruals Dummy
The Manipulation Index
?
MI =
- 2.224 + 0.221 * (Days Sales in Rec Index) + 0.102 * (Gross Margin Index) + 0.007 * (Asset Quality Index) + 0.062 * (Depreciation Index) + 0.198 * (SG&A Index) - 2.415 * (Total Accruals to Total Ast) + 0.040 * (Sales Growth Index) - 0.684 * (Abnormal Return) - 0.001 * (Time Listed) + 0.587 * (Leverage Index) + 0.421 * (Positive Accruals Dummy) - 0.413 * (Declining Cash Sales Dummy)
?
Probability of Earnings Manipulation = Normdist (MI)
Median Results
P(Manipulation)
GAAP Violators = 9.5%
Ratio Days Sales in Receivables Gross Margin Index
Control Firms = 1.1%
GAAP Violators 1.269 1.042 Control Firms 1.199 1.004
Asset Quality Index
Depreciation Index SG&A Index Total Accruals to Total Assets
0.937
0.981 0.997 0.204
0.807
1.021 0.981 0.441
Sales Growth Index
Abnormal Return Time Listed (months) Leverage Index
1.431
(0.325) 29.0 0.564
1.379
0.011 31.0 0.500
Algorithm
Beinish Probit Model
Algorithm
Algorithm
P > 9.5 Violator Firm
High Risk of Earnings manipulatio n
Probability of Earnings Manipulation
P<9.5 & P>1.1
P<1.1
Control Firm
Conclusion
90%
80%
70% 60%
Percentage of Firms
83% 76% 67%
Correctly Identified Violators (Type I Error) Incorrectly Identified Control Firms (Type II Error)
50% 40% 30% 20% 10%
45% 28.60% 20.40% 13.50% 3.60% 11.72% 5.99% 4.30% 2.94%
Cut Off Percentage
0%
Conclusion
?
Holistic Approach : Several Variables Included Inexpensive: Only 3 years of data required Simplicity
?
?
?
Including Qualitative Factors may enhance the detection power
References
?
? ?
?
?
Detecting Earnings Manipulation – A work on Benish Model. The Beneish Probit Model – Beneish1997 Simple analytical Procedures - Calderon & Green 1995 Analysis of revenue and account receivables Green & Choi www.capital9.com
doc_532379672.pptx