Accounting Restatements and the Efficiency of Industry Investment

1
Accounting Restatements and the Efficiency of
Industry Investment

• Art Durnev, Mcgill University
• Claudine Mangen, Concordia University

2
Introduction

• Wave of accounting scandals in the early 2000s
  Adelphia, Enron, Worldcom, etc.
• Various degrees of accounting errors
• accounting errors within GAAP ?
  earnings management
• accounting errors outside of GAAP ? restatements
• accounting errors outside of GAAP with intent to
  deceive ? fraud

3
Introduction

• Do accounting errors have real consequences?
• For the firms that commit the errors
• Fraud Negative announcement returns of -20
  (Palmrose et al., 2004)
• Restatements Negative announcement returns of
  between -4 and -9
• Palmrose et al., 2004 Anderson and Yohn, 2002
  General Accounting Office, 2002 Wu, 2002 Turner
  et al., 2001 Dechow et al., 1996
• Reasons
• diminished company prospects
• increased uncertainty

4
Introduction

• Competitors of firms that restate financial
  statements are adversely affected and the loss is
  significant

The number of restatements, by year Abnormal
returns of competitor firms
Mean abnormal returns are calculated over 11
event window The number of restatements for year
2002 is multiplied by two because the sample ends
in June 2002
5
The cumulative loss for competitors of restating
firms is substantial
6
Introduction

• Do accounting errors have real consequences?
• For competitors of restating firms
• Restatements Negative announcement returns of
  about -0.4 (Gleason et al., 2004 Xu et al.,
  2006)
• Why?
• This study proposes a novel explanation
• investment inefficiency

7
Investment Inefficiency Explanation

• Investments are conditioned on information about
  other companies in the industry
• Dixit and Pindyck, 1994 Admati and Pfleiderer,
  2000
• If financial statements of a firm turn out to
  have been erroneous and are restated, then
• its competitors made investments based on
  erroneous information
• its competitors' investments were inefficient
  and its competitors experiences a loss in value
  when this inefficiency is revealed

8
Aim of This Study

• Test the investment inefficiency explanation for
  competitors' loss in market value at restatement
  announcements
• Control for the explanation so far proposed in
  the literature contagion

9
Contribution

• Adds to the literature on the real effects of
  accounting irregularities for competitors
• first to propose and test the investment
  inefficiency explanation
• Extends literature on the relation between
  financial reporting and real decisions
• suggests that financial reporting has
  implications for other firms' real decisions

10
Contribution

• Contributes to the literature on the relation
  between information quality and investment
  efficiency (Durnev et al., 2004 Chen et al.,
  2006 Ferreira and Laux, 2006)
• analyzes an alternative measure of information
  quality, namely restatements
• Policy implications
• intense debate about the benefits and costs of
  the Sarbanes-Oxley Act
• the full evaluation of reforms requires
  consideration of
• the companies directly targeted by the reforms
• their competitors!

11
Hypotheses

• Information quality in a firm's financial
  statements can affect its competitors'
  investments through two channels
• 1. Indirect channel
• Firm values are correlated (Foster, 1981
  Freeman and Tse, 1992 Admati and Pfleiderer,
  2000).
• Firm's disclosure reduces information asymmetry
  about its competitors, increases liquidity and
  decreases trading costs.
• ? affects its competitors' cost of external
  financing

12
Hypotheses

• Information quality in a firm's financial
  statements can affect its competitors'
  investments through two channels
• 2. Direct channel
• Signal to its competitors when to enter a market
  (Dye, 1985).
• Signal to its competitors when to exit a market
  (Foster, 1981)
• Affects its competitors decisions regarding
  mergers and acquisitions
• Hypothesis 1. Competitors of restating firms
  experience more negative abnormal returns around
  restatement announcements when their past
  investments were more inefficient

13
Hypotheses

• Restatements occur more often in certain
  industries
• (Palmrose et al., 2004)
• - 36 in manufacturing
• - 26 in technology
• - 12 in financial and other services
• Industries with more restatements have higher
  proportion of erroneous financial statements
• Competitors in such industries have used
  increasingly erroneous information
• Hypothesis 2. In industries with more
  restatements
• past investments were more inefficient
• abnormal returns at restatement announcements
  across all competitors are more negative

14
Hypotheses

• Competitors can be defined on various SIC
  levels
• Coca-Cola versus Pepsi versus Starbucks
• Competitors on a 4-digit SIC level
• are more similar to each other than competitors
  on a 3-digit or a 2-digit level are more likely
  to take into account information about their
  peers than competitors on a 3-digit or a 2-digit
  level
• Hypothesis 3. When competitors of restating firms
  are defined at the 4-digit SIC level rather than
  the 3-digit SIC level, and when they are defined
  at the 3-digit SIC level rather than the 2-digit
  SIC level, then
• their past investments were more inefficient
• their abnormal returns at restatement
  announcements are more negative

15
Hypotheses

• Firms with a larger market are observed more
  closely
• Competitors' investments are affected more by
  financial statements of firms with a larger
  market share
• Hypothesis 4. As the market share of restating
  firms rises
• their competitors' past investments were more
  inefficient
• their competitors' abnormal returns at
  restatement announcements are more negative

16
Sample

• Restatements from Jan 1997 to June 2002 compiled
  by General Accounting Office
• search in Lexis-Nexis
• concentrates on misinterpretation of accounting
  rules and fraud
• we confirmed it
• 916 restatements in total
• 836 restatements by 785 firms remain

17
Competitors to firms that restate their financial
statements experience negative abnormal returns
around the time of the restatement announcement
18
Investment Inefficiency
19
Investment Inefficiency

• Present value of cash flows of marginal dollar
  of capital
• To measure it,
• Using 1997-2002 it can be estimated on industry
  level by pooling firms

20
Investment Inefficiency

• Measure of investment efficiency
• deviation of from optimal level
• first choose optimal level equal to 1
• corporate and dividend taxes, and other
  frictions can push optimal marginal q away from 1
• optimal q can vary across industries
• Considering taxes, optimal is around 0.8
• Try different threshold levels
• Let the data choose optimal level
• Firm estimates as deviation of from
  industry average level (Ferreira and Laux (2006))
• (1)
• (2)

21
Dividends, share repurchases, interest
Drop observations with growth in value and
capital stock gt200
22
Lower efficiency of investment (deviation of
marginal q from one) in industries with less
informative stock prices
more informative prices

• Durnev, Morck, Yeung (2004)

23
Controls

• Variables that effect competitors returns and
  industry efficiency
• industry concentration, Herfidahl index
• in more competitive industries the effect of
  restatement can be larger business conditions
  are more predictable in more concentrated
  industries
• firm and industry size
• smaller firms have larger reaction to
  restatements announcements
• leverage
• curbs agency problems of excess cash
• past performance
• market reaction to adverse news depends on
  companys past performance

24
Controls

• Variables that control for previously documented
  contagion effect
• governance quality
• competitors firms are less likely to fudge their
  numbers if they have sound governance system
• Gompers, Ishi, Metrick (2003), IRCC
• change in the dispersion of analysts' forecasts
• if restatements increase investors' uncertainty,
  then the uncertainty about competitors should
  rise around the restatement announcement
• Structural variables
• industry and firm fixed effects
• correlated errors
• omitted firm and time characteristics

25
Summary Statistics
26
H1 Competitors of restating firms experience
lower abnormal returns when their past
investments were more inefficient
27
Lagged investment efficiency of competitors in
industries with and without restatements
Longer bars indicate less efficient
investment Investment efficiency measure is
lagged by 1 year Numbers indicate t-statistics of
means comparison tests
28
Controls

• Industry size
• firms in larger industries may have more
  internal cash flow and easier access to external
  financing larger firms restate less frequently
• Diversification
• more diversified firms shift income between
  divisions more likely to restate earnings
  because of governance problems
• Leverage
• disciplining role of debt and bankruptcy costs
• Industry concentration
• Liquidity
• firms with more cash are more likely to
  overinvest
• Market-to-book
• high growth opportunities may induce investment
  uncertainty
• Industry fixed effects

29
H2 In industries with more restatements, past
investments were more inefficient
30
H2 In industries with more restatements,
abnormal returns at restatement announcements
across all competitors are more negative.
31
Over-investment and number of restatements Result
s are driven by firms that over-invest Supports
Jensens (1986) free-cash problem and Rolls
(1986) hubris hypothesis
32
H3 Comparison across different SIC levels
33
H4. As the market share of restating firms rises,
their competitors' past investments were more
inefficient
34
H4. As the market share of restating firms rises,
their competitors' abnormal returns at
restatement announcements are more negative
35
Addressing Contagion Further
Contagion may explain some of results. Two-step
regression 1. Decompose investment inefficiency
into two components part explained by
contagion part not explained by contagion 2.
Repeat using the explained and unexplained parts
of investment inefficiency
36
Sensitivity analysis

• Drop financial companies
• small fraction of the sample, results become
  stronger
• Clustered standard errors as suggested in
  Petersen (2006)
• Heteroskedastic standard errors due to industry
  averages, weighted least squares estimation
• Outliers
• Hadi (1992) method
• winsorizing of main variables
• rank-ordering of skewed variables
• Significance of abnormal returns
• logit regressions
• Issues with estimation of investment efficiency
• threshold levels of 0.75, 0.8, 0.85, 0.9
• non-linear estimation
• industry-specific thresholds
• drop firms with growth rates in value and
  capital stock gt 300, 100

37
Conclusions

• When firms announce restatements, their
  competitors experience a decline in their market
  value
• This decline occurs because competitors made
  inefficient investments based on erroneous
  financial statements
• Implication for evaluating governance
  regulations the effect of such regulation on
  competitors needs to be considered