Prior Accounting Irregularities and Market Reactions to Disclosure of Internal Control Deficiencies

1
Prior Accounting Irregularities and Market
Reactions to Disclosure of Internal Control
Deficiencies

• Ling Chen
• Department of Accounting and Information
  Management
• Universiteit Maastricht

04/04/2007
2
The SOX and the costs

• The Sarbanes-Oxley Act of 2002 (SOX)
• Section 302
• Section 404
• The first year SOX 404 implementation costs
  (excl. external audit costs).
• Firm Size Costs of revenue
• market cap. lt700 million 1.24 million 0.38
  
• market cap. gt700 million 8.51million 0.11
• source CRA International 2006

3
Motivation

• The reports on internal control systems provide
  financial statement users with an early warning
  about potential future financial statement
  problems. (PCAOB 2004)
• However, .
• Little evidence that the market reacts to firms
  disclosures of internal control deficiencies.
• No significant negative market reaction
  (Ashbaugh-skaife et al.,2005).
• Zero average excess size-adjusted returns
  (Hammersley et al., 2005).

4
Expectation

• Market only reacts to unexpected news.
• The internal control disclosure can be predicted
  by prior accounting irregularities.
• Firms without prior accounting irregularities
  experience negative market reactions.

5
Research question

• Investigate the association between internal
  control disclosures and firms prior accounting
  irregularities.
• Market reactions to internal control disclosures.

6
Sample and data

• Internal control deficiencies disclosed from
  August 2002 till December 2005 by firms listed on
  US exchanges.
• August 2002 to October 2003, hand collected from
  EDGARPlus
• November 2003 to December 2005, Compliance Week
• Irregularities data
• Restatements
• Up to June 2002, General Accounting Office
  Financial Statement Restatement Database.
• From July 2002 to December 2004, hand collected
  from the Lexis-Nexis database.
• SEC investigations, the list of Accounting and
  Auditing Enforcement Releases from the SEC
  website.
• Financial data are from Compustat.
• A matched-pairs design is used, controlling for
  size, industry, year, and stock listing.

7
Sample selection

• Number of firms disclosing ICD 987
• Firms not listed on 3 major US stock
  exchanges (218)
• Firms not covered by CRSP (5)
• Firms with missing sales data (8)
• Firms without an available matching firm (81)
• Internal control deficiency sample 675

8
Association study

• Dependent variable Expected sign
• ICD_DISCLOSURE
• Test variable
• Irregularities
• Control variables
• Financial Health (Loss)
• Complexity
• Growth
• Inventory Receivables
• Restructuring
• BIG4 /-

9
Descriptive statistics-compare means
Variable ICD Firms (n675) Control Firms (n675) Difference
Restatement 10.96 3.56 7.41
SEC investigations 2.07 0.44 1.63
Loss Indicator 40.44 29.78 10.67
Segments 2.231 1.967 0.265
Foreign Sales 43.26 42.96 0.30
Growth 31.70 32.59 -0.89
Inventory Receivables 548.954 503.694 45.260
Restructuring -0.016 -0.011 -0.006
BIG4 0.828 0.831 0.003
, , indicates significance at the .01,
.05, and .10 level or better, based on
t-statistic for difference in means.
10
Results
Logistic regression DVICD_DISCLOSURE Logistic regression DVICD_DISCLOSURE Logistic regression DVICD_DISCLOSURE Logistic regression DVICD_DISCLOSURE Logistic regression DVICD_DISCLOSURE
Variable Exp Sign Model 1 Model 2 Model 3
Intercept /- -0.577 -0.670 -0.665
Irregularities 1.248
Restatements 1.151
AAER 1.369
Loss Indicator 0.503 0.488 0.487
Segments 0.131 0.132 0.132
Foreign Sales -0.141 -0.167 -0.163
Growth -0.009 0.033 0.036
Inventory Receivables 0.387 0.442 0.439
Restructuring -1.686 -1.519 -1.512
BIG4 /- 0.071 0.057 0.055
Likelihood ratio Chi-Square 33.35 66.07 65.39
, , indicates significance at the .01,
.05, and .10 level or better, based on Wald
chi-square statistics
11
Other analyses (contd)
Logistic regression DVICD_DISCLOSURE Logistic regression DVICD_DISCLOSURE Logistic regression DVICD_DISCLOSURE Logistic regression DVICD_DISCLOSURE Logistic regression DVICD_DISCLOSURE
Variable Exp Sign Ordered Logistic (n1350) MW (n1014) ICDSD (n336)
Intercept /- -0.734 -0.464
Irregularities 1.139 1.280 1.198
Loss Indicator 0.523 0.595 0.162
Segments 0.094 0.125 0.141
Foreign Sales -0.136 -0.231 0.012
Growth 0.040 0.099 -0.112
Inventory Receivables 0.463 0.317 0.745
Restructuring -1.604 -1.076 -3.093
BIG4 /- 0.104 0.148 -0.204
Likelihood ratio Chi-Square 68.61 57.38 14.67
, , indicates significance at the .01,
.05, and .10 level or better, based on Wald
chi-square statistics.
12
Findings

• Prior accounting irregularities represented by
  financial statement restatements and SEC
  investigations are found highly significant in
  predicting internal control deficiency
  disclosures.
• Question whether the market reaction to ICD
  disclosures differs between firms with and
  without prior accounting irregularities?

13
Market reactions to internal control deficiency
disclosures

• Event study

is the return of the sample firm i on day t is
the corresponding size decile portfolio return
from CRSP on day t and is the number of sample
stocks whose abnormal returns are available at
day t.
14
Market reaction study

• Dependent variable Expected sign
• CAR
• Test variable
• Irregularities
• Other independent variables
• Severity of ICD /-
• SOX404 /-
• BIG4 /-
• Other news
• Earnings surprise
• Type of filings /-
• Prediction indicator

15
Market reactions to ICD disclosures, univariate
results
CAR (-1,1) Mean t-statistics Negative
All disclosures (n643) 0.17 0.562 49.14
Uncontaminated (n85) -2.04 -2.850 52.94
Contaminated (n558) 0.51 1.502 48.57
Difference between uncontaminated and contaminated disclosures -2.55 -2.777 -2.125
CAR (-1,4)      
All disclosures (n643) 0.55 1.583 49.77
Uncontaminated (n85) -1.83 -2.107 60
Contaminated (n558) 0.915 2.426 48.21
Difference between uncontaminated and contaminated disclosures -2.74 -2.683 -2.309
, , indicates significance at the .01,
.05, and .10 level or better, based on
t-statistic for difference in means or
z-statistics for Wilcoxon sign-rank tests.
Uncontaminated refers to the group of internal
control disclosures without other material news
contemporaneously released and contaminated
refers to the group of internal control
disclosures with other material news
contemporaneously released.
16
Market reactions to uncontaminated ICD
disclosures by prior accounting irregularities,
univariate results
CAR (-1,1) Mean t-statistics Negative
Firms without prior accounting irregularities -2.24 -2.886 51.95
-2.24 -2.886 51.95
Firms with prior accounting irregularities -0.07 -0.057 62.50
-0.07 -0.057 62.50
Difference between without and with prior accounting irregularities -2.17 -0.886 -0.572
CAR (-1,4)      
Firms without prior accounting irregularities -2.40 -2.919 59.74
-2.40 -2.919 59.74
Firms with prior accounting irregularities 3.67 0.806 62.50
3.67 0.806 62.50
Difference between without and with prior accounting irregularities -6.07 2.079 -0.828
, , indicates significance at the .01,
.05, and .10 level or better, based on
t-statistic for difference in means or
z-statistics for Wilcoxon sign-rank tests.
Uncontaminated refers to the group of internal
control disclosures without other material news
contemporaneously released
17
Multivariate results
    Dependent VariableCAR (n422) Dependent VariableCAR (n422)
Independent Variables Predicted Sign CAR (-1,1) CAR (-1,4)
Intercept   -0.023 -0.027
Intercept  
Irregularities 0.019 0.028
Irregularities
MW 0.011 0.015

SOX404 -0.011 -0.008

BIG4 0.001 -0.001

Filing10KQ 0.026 0.028
   
Adjusted R2   0.04 0.03
Form10-K/Q disclosures with early earnings
announcements and uncontaminated Form 8-K
disclosures
18
Additional analysis I
    Dependent VariableCAR (n422) Dependent VariableCAR (n422)
Independent Variables Predicted Sign CAR (-1,1) CAR (-1,4)
Intercept   -0.022 -0.025
Intercept  
Predict 0.017 0.025
Predict
MW 0.010 0.014

SOX404 -0.011 -0.008

BIG4 0.001 -0.001

Filing10KQ 0.025 0.027
   
Adjusted R2   0.04 0.03
19
Additional analysis II
    Dependent VariableCAR (n422) Dependent VariableCAR (n422)
Independent Variables Predicted Sign CAR (-1,1) CAR (-1,4)
Intercept   -0.021 -0.029
Intercept  
Irregularities 0.021 0.029
Irregularities
Loss Indicator -0.011 -0.007

Segments 0 0.002

MW 0.013 0.016

SOX404 -0.012 -0.008

BIG4 0 -0.002

Filing10KQ 0.027 0.029
   
Adjusted R2   0.04 0.03
20
Additional analysis III
    Dependent VariableCAR (n422) Dependent VariableCAR (n422)
Independent Variables Predicted Sign CAR (-1,1) CAR (-1,4)
Intercept -0.012 -0.009
Irreg_404 0.003 0.01
Nonirreg_404 - -0.023 -0.027
Nonirreg_302 - -0.01 -0.017
MW 0.01 0.014
BIG4 0.001 0
Filing10KQ 0.026 0.028
Adjusted R2   0.04 0.03
21
Conclusion

• The market reacts less negatively to ICDs
  disclosed by firms with prior accounting
  irregularities than to disclosures made by firms
  without.
• Explains the lack of significant market reactions
  to ICD disclosures is partly due to the
  predictability of ICD disclosures by prior
  accounting irregularities.