You are on page 1of 20

Journal of Accounting and Economics 6 (1984) 185-204.

North-Holland

ACCOUNTING ACTIVITIES, SECURITY PRICES, AND CLASS


ACTION LAWSUITS

Robert L. KELLOGG*
Southern Methodist University, Dallas, TX 75275, USA

Received October 1981, final version received October 1984

Provisions in the securities acts provide incentives to purchasers of common stocks to initiate class
action lawsuits when stock prices decline at and preceding announcements that directly reduce, or
imply a reduction in, previously reported accounting book values. Reported common stock returns
associated with alleged misrepresentations in financial statements are consistent with incentives
provided by the law. Classification of misrepresentations based on hypothesized relations between
announcements and security returns results in observed differences in the association between
litigated accounting announcements and common stock returns.

I. Introduction
Expected legal costs, including costs of class action lawsuits under the
federal securities acts, are thought to be an important variable in the market
for audited financial statements)Although the legal cost variable has been
incorporated in models of accounting decision-making, the only empirical
evidence of the nature of the function relating legal costs to accounting
activities has been presented in subjective terms culled from statutes and
reported judicial decisions. 2 Class action lawsuits against preparers and audi-
tors of financial statements are based primarily on sections of the securities
acts that explicitly condition liability on security prices and their relationships
to accounting announcements. This study analyzes observed associations be-
tween stock prices and accounting announcements in light of these legal
provisions.
In the next section, analysis of the sections of the securities acts most
frequently used to assert class action liability for accounting misrepresenta-
tions (activities alleged to have misled investors) derives two predictions about
the relationship between misrepresentations and stock returns: (1) abnormal
* The timely and incisive comments of George Foster contributed greatly to the completion of
this study.
1See, e.g., Antic (1982) and Wallace (1980).
2St.Pierre and Anderson (1984) reference interpretive studies and present their own analysis
based on published reports.

0165-4101/84/$3.001984, Elsevier Science Publishers B.V. (North-Holland)


186 R . L Kellogg, Class action lawsuits

security returns will be observed when misrepresentations are discovered, and


(2) abnormal security returns also will be observed in periods preceding
discovery. In section 3, observed returns in a sample of lawsuits are found to
be consistent with these predictions. Section 4 presents returns series associated
with litigated accounting announcements classified according to their expected
relationship to stock prices. The results reported in sections 3 and 4 are
summarized in section 5 and related to the importance of accounting research
to the legal community.

2. The law, stock returns, and misrepresentations


An important source of expected legal costs to preparers and auditors of
corporate financial statements is the threat of class action lawsuits brought
under the federal securities acts. Most frequently, all persons who purchased a
firm's security during the period that a misrepresentation remained uncor-
rected make a claim to be indemnified for any trading or holding losses on the
security) Although the law applies to all classes of securities, because of the
availability of common stock returns data, this study will consider only
lawsuits by common stock purchasers. I will refer to these as 'accounting
buyers' suits'.
The law of accounting buyers' suits makes stock returns at and preceding
public discoveries of misrepresentations important to understanding the legal
liability associated with the preparation and certification of financial state-
ments. Nearly all accounting buyers' suits are brought under Section 11 of the
Securities Act a n d / o r Rule 10b-5 of the Securities and Exchange Commission
(SEC). 4 Section 11 applies specifically to information included in a registration
statement filed with the SEC, and Rule 10b-5 applies generally to purchases
and sales of securities. Both provisions contain similar language establishing
civil liability for making 'an untrue statement of a material fact' or omitting
' t o state a material fact necessary in order to make the statements made not
misleading'. Section 11 and Rule 10b-5 contain causation requirements and
damage measures that suggest that negative stock returns will be observed
when misrepresentations are discovered and in preceding periods of time.

2.1. Causation and returns at discoveries

Both Section 11 and Rule 10b-5 require causal relationships between misrep-
resentations and damages, although the burden of proof differs. Under Section

3 The sampling procedures described in Section 2 produced 106 lawsuits by classes defined as
common stock purchasers and 14 by sellers of common stock.
415 U.S.C. 78a-jj (1970) and 17 C.F.R. 240.10b-5 (1980). These sections of the law are cited
in 105 out of 106 lawsuits identified for this study.
R.L Keliogg~ Class action lawsuits 187

11, defendants can avoid liability for any stock price declines that they can
prove were caused by something other than the misrepresentations with which
they are charged. Under Rule 10b-5, plaintiffs must prove that defendants'
misrepresentations caused the stock price declines that constituted plaintiffs'
damages.
Plaintiffs in an accounting buyers' suit meet the causation requirement by
pleading that stock prices conditioned by 'false' accounting data (those pub-
fished prior to misrepresentation discovery) were higher than those conditioned
by 'true' accounting data. True accounting data are those announced when a
misrepresentation is corrected or those implied when discovery of a misrepre-
sentation occurs other than by explicit correction of accounting data. 5 In
effect, plaintiffs are claiming: 'Had we but known these true accounting data,
we would not have paid what we did for these securities.' A rapid stock price
adjustment at discovery seems critical to this legal theory.

2.2. Damage measures and returns preceding discoveries


The damages to be recovered by each plaintiff in an accounting buyers' suit
are based on stock prices. Under Section 11, damages are the difference
between the price at which the security was purchased and either (a) the price
at which the security was sold, or (b) the price of the security at the date the
lawsuit was filed. 6 Under Rule 10b-5, a plaintiff's damages are the difference
between the purchase price and either (a) the price at which the security was
sold or (b) the price of the security when the market has adjusted to public
discovery of the alleged misrepresentation.
In cases involving actively traded securities, the courts have regarded stock
price declines after market adjustment to public discovery of a misrepresenta-
tion as not caused by misrepresentation. 7 This has the effect of making the
legal damage measure in both Section 11 and 10b-5 cases the difference
between purchase price and either selling price or the price immediately after a
misrepresentation is publicly discovered. Purchases and sales months (in some
cases, years) before discovery of a misrepresentation can be used in computing
the damages to be received by plaintiffs and the contingent fees of their
lawyers. This provides incentives to prosecute cases in which negative returns
precede discovery. Observed returns to stockholders of firms charged with

5Discovery can be through correction (the announcement of revised accounting numbers) or


through less explicit means such as announcement that earnings were 'grossly overstated' or that
an allowance account would be increased. All discoveries either decrease firms' book values or
imply that such a decrease will occur.
6Section ll(e). There is a three-year (from the date of a security offering) statute of limitations
on Section 11 actions.
7Harris v. American Investment Company, 523 F. 2d 220 (8th Cir. 1975). Esplin v. I-Iirschi, 402
F. 2d 94 (10th Cir., 1968). Chasins v. Smith, Barney & Co., 306 F. Supp. 177 (SDNY, 1969).
188 R.L Kellogg, Class action lawsuits

issuing misleading accounting data may be negative prior to discovery solely


due to these incentives built into the damage measure; prediscovery negative
returns increase the probability that an accounting activity will result in a
lawsuit. This implies that ex post selection bias affects observed returns for any
sample selected on the basis of involvement in accounting buyers' suits.
Even in the absence of ex post selection bias, prediscovery negative returns
could be predicted on the basis of previous empirical results. Nonlitigated
accounting announcements (e.g., reduced earnings) that were predicted to be
associated with decreases in firms' market values have been preceded by
abnormal negative security returns. 8 These results have been construed as
evidence that capital markets obtain much of the potential information in
accounting numbers from more timely non-accounting sources. Some have
theorized, however, that these alternative information sources are ineffective in
cases of accounting misrepresentations. Chambers (1974, p. 48) claims that
'even though the market may readily adjust to information which becomes
available, it has no ready means of countering.., false, misleading and erro-
neous numbers.' Anderson (1977, p. 421) concludes that 'in the major liability
cases to date the market has been fooled by the failure of the reporting system
to make timely disclosure of relevant information . . . . This is not too surpris-
ing, because there are certain types of information for which it is difficult to.
conceive of alternative sources.' These writers seem to imply the existence of a
discrete decrease in security prices when a misrepresentation is discovered.
Returns at and preceding dates of discoveries of misrepresentations that
have resulted in accounting buyers' suits are reported in the next two sections.
The extent to which observed returns reflect the incentives built into the
securities laws or the effects of other more general factors can be estimated
only by comparing returns associated with litigated and non-litigated account-
ing activities. The data analyzed in this paper relate to only litigated account-
ing activities, and only descriptive results are presented.

3. Observed stock returns at and preceding discoveries of misrepresentation

3.1. The sample


The preceding section described elements of the securities laws that provide
incentives to initiate a buyers' accounting suit in response to an announcement
that affects financial statements when: (a) the announcement is associated with
negative stock returns and (b) the firm's common stockholders have earned
negative returns prior to the announcement. In this section, risk-adjusted
common stock returns associated with a sample of accounting buyers' suits are

SSee, e.g., Ball and Brown (1968), Brown (1970), Foster (1975).
R.L. Kellogg, Class action lawsuits 189

reported. Accounting buyers' class actions were identified from:

(1) all stockholder lawsuits reported in firms' filings with the SEC between
1967 and 1976 as reported in the Disclosure Journal.
(2) all cases filed in the United States District Court for the Southern District
of New York between 1971 and 1976 that named large accounting firms
(the 20 largest) as defendants. 9
(3) Wall Street Journal reports of lawsuits against large accounting firms (the
20 largest).
(4) litigation footnotes in firms' 1976-1979 annual reports listed on Mead
Data's National Automated Accounting Research System.
(5) reported opinions of the federal courts that included the words '10b-5' and
'account' and were listed on Mead Data's LEXIS system.
(6) lawsuits mentioned in previous studies of accountants' liability.

The litigation sample is limited to one lawsuit involving the financial state-
ments of any given corporation. Frequently, there are several lawsuits initiated
roughly contemporaneously against a single corporate defendant, each case
involving a separate class of stock purchasers, defined by dates of purchase or
the federal district court in which they are being represented. In these instances,
the earliest lawsuit filed among those for which the class of plaintiffs could be
determined was included in the sample.
One hundred and six accounting buyers' class actions, each involving the
financial statements of a different corporation, were identified in which particu-
lar accounting practices were alleged to have misled common stock purchasers.
The stock returns analysis to be performed on the sample requires available
returns data and a discovery date for each firm. Seventy-six of the 106 firms
named in these lawsuits were listed on the daily security price tapes published
by the Center for Research on Security Prices (CRSP) for periods of at least six
months prior to the filings of the class action complaints based on their
financial statements. It was possible, from the sources available, to identify the
alleged dates of discovery in cases against 72 of these 76 firms.
The period over which damages are to be assessed (the Damage Computa-
tion Period) in buyers' class actions involving publicly traded securities is
defined by the courts as extending from the date on which a misrepresentation

9A list of reported decisions in lawsuits against accountants prepared by a national accounting


firm reported 89 of 180 decisions by federal district courts written by the Southern District of New
York. The next most active court (the Eastern District of Pennsylvania) accounted for 13 decisions.
Another samphng source, the Disclosure Journal ceased publication in 1976. The fact that the
sample is composed primarily of pre-1977 lawsuits is not as limiting as it might appear. Since the
Hochfelder decision by the Supreme Court in 1976 (425 U.S. 185) there have been fewer
accounting buyers suits as the federal courts have attempted to clarify the apparently more
stringent standards to be met by plaintiffs under Hochfelder. [See Sanders v. John Nuveen & Co.,
Inc., 554 F. 2d 790 (7th Cir. 1977) and McLean v. Alexander, 599 F. 2d 1190 (3rd Cir. 1979).]
190 R.L. Kellogg, Class action lawsuits

occurs through the date on which the market (not an individual plaintiff)
learns of the existence of the misrepresentation (the discovery date). Security
returns during the damage computation period are directly related to stock
price changes that are used to compute the legal damages that are associated
with accounting misrepresentations. The returns analysis in the next section
focuses on this period of time.
Legal complaints that are filed in class action lawsuits must include a
statement of the damage computation period; it defines the class of plaintiffs.
Unfortunately, the periods stated in complaints cannot be relied upon to
identify the discovery date. It is in plaintiffs' interest to name a long damage
computation period and often damages are claimed not to the discovery date,
but to the date on which a lawsuit is filed. The Wall Street Journal lndex, firms'
financial statements, and SEC filings were searched to identify public dis-
closures of the existence of misrepresentations made prior to the dates of
discovery alleged by plaintiffs This resulted in a unique damage computation
period for each of 72 CRSP firms identified as either (1) the period over which
damages were claimed, or (2) the period from the date of alleged misrepresen-
tation occurrence through the date on which public disclosure first was made
of the existence of a misrepresentation. Analysis of returns data focuses on
these damage computation periods.
Returns analysis uses a market model estimation period (described below)
preceding each firm's discovery date. The estimation period is limited to
periods preceding the event of interest (discovery) by severe sample deteriora-
tion at and following discovery dates due to trading suspensions and delistings.
Monthly returns during damage computation periods and daily returns during
months of discovery are analyzed. To insure reasonable estimation periods, 56
of the 72 CRSP firms meet at least one of the following two requirements:

(1) Continuous stock returns data are available for at least 24 months prior to
the damage computation period and through the damage computation
period. This provides at least a 24-month market model estimation period
and returns through the damage computation period. 43 firms meet this
requirement.
(2) Continuous stock returns data are available from 220 days prior to the day
of misrepresentation discovery up to two days before the day of discovery.
This provides a 200-day market model estimation period and at least 18
days of returns. 48 firms meet this requirement. 1

Return statistics are reported for the two intersecting subsets of the 56 firms
meeting at least one requirement. The firms are listed in an appendix with

lo Only 42 of 48 firms had daily returns available through the day of discovery. The remaining
six firms were included to provide the largest possible sample for generating return statistics for
days close to discovery.
R.L Kellogg, Class action lawsuits 191

industry classification (SIC code), the lengths (in months) of associated damage
computation periods and market capitalization (stock price times number of
shares outstanding) at the beginnings and ends of damage computation peri-
ods. 11

3.2. S t o c k return statistics

Causation and damage elements of the law of accounting buyers' suits


suggest the existence of negative returns at dates of discovery of misrepresenta-
tions and in preceding time periods. Two null hypotheses are tested, stated in
terms of market model prediction errors:

Null 1. the mean market model prediction error at dates (months or days) of
discovery (MPEo) equals zero, and
Null 2. the mean cumulative prediction error through months immediately
preceding months of discovery (MCPE n _ 1) equals zero.

Market model prediction errors are calculated using estimated coefficients


from the market model:

where
Rj~ t = return to firm j common stock in period t,
~, b = estimated coefficients, and
R m , t = NYSE value weighted index in period t. 12

Hypotheses are tested using the z-statistic described by Pattell (1976).

3.3. Observed return statistics


Two return series are analyzed:
(a) monthly returns over firms' damage computation periods for the 43 firms
meeting the monthly data requirements discussed above in describing the
sample, and
(b) daily returns for the 48 firms meeting the daily data requirements from 20
days before discovery through 10 days after discovery.

Prediction error and cumulative prediction error statistics for each of the
two observation periods are reported in tables 1 and 2, respectively. In

11The appendix includes a classification of firms by the predominant type of misrepresentation


alleged by lawsuit plaintiffs. The basis for classification is described in section 4.
12'Periods' are discussed rather than months or days because procedures are performed using
both monthly and daily returns.
192 R. L Kellog& Class action lawsuits

Table 1
Mean monthly prediction errors (MPE) for the last 25 months of damage computation periods
with mean cumulative prediction errors (MCPE) for 43 firms with monthly data.
Month
( D = discovery
month) M P E ( + , - )a Z M C P E ( + , - )a Z

D-24 -0.036 (5,11) -1.32 - 0 . 0 6 5 (3,13) -4.02


D-23 -0.043 (6,10) -0.90 - 0 . 1 5 0 (3,13) -4.20
D-22 0.026 (6,11) 2.09 - 0 . 0 8 5 (4,13) -3.54
D - 21 -0.024 (7,11) -0.50 - 0 . 1 2 0 (5,13) -3.30
D - 20 -0.064 (5,14) -2.66 - 0 . 0 9 3 (3,16) -3.16
D - 19 - 0.036 (8,12) - 1.54 - 0.070 (3,17) - 4.57
D-18 -0.025 (7,13) -1.72 - 0 . 0 7 3 (4,16) -4.97
D - 17 -0.052 (5,17) -2.78 -0.001 (4,18) -5.75
D-16 -0.015 (8,14) -0.54 0.013 (4,18) -5.68
D - 15 -0.051 (8,17) -4.86 - 0 . 0 2 4 (7,18) -5.71
D-14 -0.052 (5,21) -2.87 - 0 . 1 3 2 (6,19) -6.38
D - 13 -0.005 (14,13) 0.59 - 0 . 1 7 5 (5,23) -5.96
D - 12 -0.046 (13,16) -2.85 - 0 . 2 2 9 (6,23) -6.65
D - 11 -0.032 (11,19) -0.85 - 0 . 2 1 9 (7,23) -6.63
D - 10 0.016 (20,11) 0.58 - 0 . 2 0 4 (9,22) -5.89
D- 9 -0.024 (16,17) -1.81 - 0 . 1 4 6 (9,24) -6.14
D- 8 -0.017 (9,24) -2.39 - 0 . 2 0 5 (10,23) -6.99
D - 7 - 0.047 (10, 26) - 3.45 - 0.234 (8, 28) - 6.75
D - 6 - 0.059 (14, 22) - 2.84 - 0.241 (10, 26) - 7.25
D - 5 - 0.069 (10,28) - 4.79 - 0.264 (5, 33) - 8.90
D- 4 -0.009 (11,29) 0.32 - 0 . 2 3 8 (7,33) -5.37
D- 3 -0.045 (13,27) -4.31 - 0 . 2 5 7 (7,33) -6.97
D - 2 - 0.032 (13, 30) - 1.60 - 0.279 (7, 36) - 7.45
D- 1 -0.063 (13,30) -3.46 - 0 . 3 1 7 (7,36) -8.63
D -0.084 (13,30) -5.27 - 0 . 4 2 7 (6,37) -10.10

Average market model coefficients: A -- 0.004 B = 1.39 R 2 = 0.21


Average damage computation period = 25.6 months

a Number of cases with positive prediction errors and number of cases with negative prediction
errors in parentheses.

discussing the first null hypothesis (MPE n = 0), I will refer to the left-hand
column of each of these tables; and in discussing the second null hypothesis
( M C P E n - 1 = 0), I will refer to the right-hand column of table 1.
Mean monthly prediction errors (MPE) are cross-sectional averages. In
calculating the mean monthly cumulative prediction errors ( M C P E ) reported
in table 1, prediction errors are cumulated in a way that differs from the
standard approach and that results in only returns for which plaintiffs claim
indemnification being included in the reported averages. This makes possible a
comparison of M C P E o _ t with MCPE n that provides an estimate of the
average percentage of the returns on which plaintiffs base their claims that
occurs t months or more before a misrepresentation is discovered. The mean
cumulative prediction errors are average 'abnormal performance indices' or
R . L . Kellogg, Class action lawsuits 193

Table 2
Mean daily prediction errors (MPE) and mean cumulative prediction errors ( M C P E ) from 20
days before through 10 days after the day of discovery (D) for 48 firms with daily data.
Day MPE ( +, - ) Z MCPE ( +, - ) Z

D - 20 - 0.007 (23, 25) - 1.33 - 0.007 (23, 25) - 1.33


D - 19 0.004 (28, 20) 0.83 - 0.002 (24, 24) - 0.35
D - 18 0.014 (22,26) 1.27 0.001 (28,20) 0.45
D-17 -0.005 (18,30) -0.35 0.006 (20,28) 0.21
D - 16 0.002 (27, 21) 0.60 0.006 (23, 25) 0.46
D - 15 0.012 (28, 20) 2.03 0.018 (27, 21) 1.25
D-14 0.000 (24,24) 0.33 0.019 (23,25) 1.28
D - 13 -0.008 (14,34) -2.26 0.011 (21,27) 0.40
D - 12 0.000 (24,24) 0.11 0.009 (27,21) 0.41
D - 11 - 0.002 (23, 25) - 0.30 0.006 (24, 24) 0.29
D - 10 0.002 (26,22) 0.79 0.008 (27,21) 0.52
D - 9 - 0.005 (18, 30) - 1.00 0.004 (26, 22) 0.21
D - 8 0.006 (29,19) 1.37 0.009 (23, 25) 0.58
D - 7 0.002 (20, 28) - 0.46 0.011 (24, 24) 0.44
D-6 0.003 (20,28) -0.15 0.008 (25,23) 0.38
D - 5 0.012 (17, 31) - 1.84 - 0.004 (23, 25) - 0.09
D-4 0.020 (28,20) 3.30 0.017 (24,24) 0.72
D- 3 -0.002 (21,27) -0.99 0.016 (22,26) 0.46
D - 2 - 0.011 (16, 32) - 1.93 0.005 (22, 26) 0.01
D - 1 - 0.024 (22, 24) - 5.20 - 0.032 (18, 28) - 1.56
D - 0.039 (15, 27) - 8.38 - 0.075 (13, 29) - 3.35
D + 1 0.009 (18,18) 1.78 - 0.074 (10, 26) - 2.44
D + 2 -0.005 (17,20) -3.17 -0.082 (9,28) -3.23
D+ 3 0.008 (16,22) -0.93 - 0 . 0 9 2 (11,27) -3.38
D + 4 0.000 (20,17) 0.21 - 0.089 (10, 27) - 2.98
D + 5 0.003 (21,17) 0.34 - 0.081 (10, 28) - 2.68
D + 6 0.001 (17, 21) - 0.59 - 0.080 (11, 27) - 2.74
D + 7 -0.021 (20,19) -1.68 - 0 . 1 0 0 (11,28) -2.89
D + 8 0.003 (21,17) 0.40 - 0 . 0 7 4 (13,25) - 2.38
D + 9 0.001 (21,16) 0.14 - 0.075 (12, 25) - 2.38
D + 10 -0.001 (18,20) -0.89 - 0 . 0 7 8 (15,23) -2.86

Average market model coefficients: A = -0.002 B = 1.45 R 2 = 0.34

aSee footnote to table 1.

API [see Ball and Brown (1968)]. They represent the average percentage
declines in firms' stock prices from the beginnings of damage computation
periods. Since each lawsuit involves a unique damage computation period, the
individual abnormal performance indices that are being averaged include
differing numbers of months. If a firm's damage computation period is T + 1
months long, both M P E o _ t and M C P E o _ t contain the prediction error from
the first month of the damage computation period for that firm. For example,
the averaging that results in MCPED_ 6 includes a 76-month API for Genesco
(82-month damage computation period) and a 2-month API for Sanitas Service
(8-month damage computation period); it does not include an API for Kleinerts
(6-month damage computation period). Table 1 reports average monthly
194 R.L Kellogg, Class action lawsuits

prediction errors over 25 months including discovery, a period approximating


the average damage computation period for the sampled lawsuits. Sixteen firms
had damage computation periods of more than 25 months; thus, M C P E n _ 2 4
reflects returns to stocks of these 16 firms in months preceding D - 24.
Mean prediction errors at discovery (MPED) are significantly less than zero
(with less than a 5% probability of Type 1 error) for the monthly and the daily
returns series. For the 43 firms for which prediction errors could be calculated
over entire damage computation periods (table 1), M P E D = - 0 . 0 8 4 and
Z D = - 5.27. Table 2 reports M P E n = - 0.039 and Z = - 8.38 for the 42 firms
for which prediction errors on the day of discovery were calculated. Significant
negative returns are also evidenced during the two days immediately preceding
discovery. 13
The results provide evidence that negative returns are associated with
discoveries of accounting misrepresentations. This can be regarded as con-
sistent with plaintiffs' theory that security prices conditioned by accounting
data prior to discovery were greater than prices conditioned by accounting
data incorporating misrepresentation discovery./4 Fischel (1982) has argued
that significant market model prediction errors at discovery should be a sine
qua non for recovery by plaintiffs. 15 The results indicate that, on average,
plaintiffs in these cases have met that requirement.
Mean cumulative prediction errors prior to discovery are significantly less
than zero. During damage computations periods (table 1, right-hand columns),
this is true as much as two years before discovery: MCPED_24 = 6.5% and
Z = - 4 . 0 2 . A comparison of M C P E n_ t with M C P E D provides an estimate of
the average percentage of the returns on which plaintiffs base their claims that
occurs t months or more before a misrepresentation is discovered. On average,
more than one-half of the negative returns for which plaintiffs seek indemnifi-
cation occurs more than one year before a misrepresentation is discovered:
M C P E n - 12 = - 22.9% and M C P E D = - 42.7%. By the beginning of the month
of discovery, on average, about 3 / 4 of the total market-adjusted negative
return over the damage computation period has occurred: M C P E o _ 1 =

13Incorporation of lead and lag factors in daily market model regressions in order to take into
account non-synchronous trading [see Scholes and Williams (1977)] also produced statistically
significant negative prediction errors on days of discovery.
14Note, however, that, given the existence of legal liability for accounting misrepresentations,
the validity of this theory does not depend upon accounting data's being a timely source of
information about firms' operations. Purchasers of securities prior to discovery are eligible to be
included as plaintiffs in accounting buyers' suits, while purchasers subsequent to discoveries of
misrepresentations are not. This means that post-discovery security prices are analogous to
ex-dividend prices; transactions subsequent to discoveries involved securities that lack an economi-
cally valuable right that is included in pre-discovery trades.
15Fischell (1982, p. 17) states: 'The market model makes it possible to test whether false
information caused a security to trade at an artificially high or low price by measuring whether
investors earned any abnormal returns at the time the correct information was released to the
public.'
R.L Kellogg Class action lawsuits 195

- 313% and M C P E o = - 42.7%. 36 of the 43 firms had negative returns prior


to the months in which misrepresentations were discovered.
These observed pre-discovery abnormal returns are consistent with the
incentives provided by the federal securities laws. All 56 lawsuits for which
associated returns are reported are subject to the incentives built into the law. 16
But it is well documented that not all accounting announcements exhibit the
same association with stock returns. 17 In the next section, the 56 lawsuits are
classified according to the type of information announced at discovery so that
possible differences in returns during damage computation periods can be
examined.

4. Returns associated with lawsuits classified by information announced


at discovery
The information made pubhc at discovery in the cases for which returns data
are available related to many different financial statement items and specific
amounts; but the information announced in each case can be classified as one
of four types:

Type 1: prior financial statements were affected by fraud,


Type 2: prior financial statements were affected by information processing
mistakes,
Type 3: prior financial statements failed to separately disclose certain trans-
actions, or
Type 4: realizable values of assets reported in prior financial statements are
being reduced.

The type of information announced at discovery associated with each firm in


the sample is noted in the appendix. The use of the term 'fraud' should not be
construed as implying illegal or dishonest acts by defendants in the lawsuits so
classified. These cases (Type 1) involved allegedly intentional acts by someone
(in some cases, agents external to the firms directly involved in the accounting
buyers' suits) that affected the financial statements with which defendants were
associated. The Type 3 omitted transactions involved corporate insiders in
most cases; in all cases plaintiffs allege that transaction occurrence should
have been reported in footnotes. The assets revalued in the Type 4 cases were

16Six of the 56 cases exhibited positive predictions errors over the damage computation period.
Unadjusted returns also were positive in five of these cases. It is difficult to understand plaintiffs'
incentives to invest in these lawsuits. In one of these cases (Mohawk Data) the court decided in
favor of plaintiffs and awarded zero damages.
X7See Foster (1980) and Kaplan (1978) for reviews of studies associating security returns with
accounting announcements.

J.A.E.-- B
196 R.L. Kellogg. Class action lawsuits

receivables, inventories, capitalized product development costs, investments in


subsidiaries, and marketable securities.
The fourth type of information (revision of asset realizable values) is
strikingly different from the first three. There is no implication that prior
financial statements were at the time of their publication imperfect or incom-
plete. Realizable values regularly change between financial statement prepara-
tion dates. This distinguishing characteristic of Type 4 information has been
recognized by both the accounting and the legal communities. Carmichael
(1976) calls these 'normal recurring adjustments' and cites as examples loan
loss reserves and lower of cost or market inventory valuations. Kripke (1971)
terms them 'probabilities' and cites accounts receivable, inventories, and fixed
asset values based on future service potential. Simunic (1980, p. 173) dis-
tinguishes inventories and receivables from other accounting valuations be-
cause 'the valuation of these items is a complex task requiring a forecast of
future events'.
The federal courts repeatedly have recognized the special character of Type
4 information when liability under the securities laws is at issue. They
emphasize the subjective nature of estimated valuations and their susceptibility
to changing conditions:

The complaint here asserts an overstatement of earnings by C.I. due to


failure to disclose the need for, and make, proper reserve for anticipated
losses... [T]he complaint charges only that the accounting judgment of
plaintiff in this fluid situation would have been different from the account
judgment of Peat Marwick. 18
Plaintiff's complaints about the annual reports all revolve around the
dates of reporting of what the plaintiff terms 'anticipated simulator losses'
and the accuracy of the accounting judgments made. As to the latter, these
are matters of judgment, a judgment dearly dependent upon constantly
changing factors. 19
As one leading commentator points out, many of the 'facts' which cases
treat as material, such as the value of accounts receivable or even the value
of fixed assets, are essentially probabilities as to future capacities or
expectations. 2

It is reasonable to expect the probabilistic and forward-looking nature of


Type 4 valuations to be recognized in capital markets. There is some support
for a prediction that changes in realivable values will, on average, be associated

XSLewis v. Black, CCH Fed Sec. L. Rep. P95, 312 (EDNY, 1976) at pages 90, 167.
19polin v. Conductron Corp., 552 F. 2d 797 (8th Cir. 1977) at page 806.
2Marx v. Computer Sciences Corp. 507 F. 2d 458 (9th Cir. 1974) at note 5.
R.L. Keliog~ Class action lawsuits 197

with less concentration of stock price declines in the discovery month. 2t


Market model prediction errors are analyzed to investigate whether the distinc-
tion between changes in realizable values and other types of misrepresentations
that is noted by accountants, lawyers, and the courts is reflected security prices.
In tables 3 and 4, prediction errors are reported in panel A for changes in
realizable values and in panel B for fraud, mistake, and omitted transactions.
Mean monthly and daily prediction errors ( M P E ) are reported in the
left-hand columns of tables 3 and 4, respectively. The month of discovery
(table 3) is associated with negative prediction errors for both changes in
realizable values (MPED= -4.8% in panel A) and for fraud, mistake, and
omitted transactions (MPE D = -10.8% in panel B). The observed prediction
errors are statistically significant for fraud, mistake, and omitted transactions
(z = -5.67). Daily prediction errors (table 4) are, on average, significantly
negative at discovery of both classes of misrepresentations. Since public
sources of corporate announcements (including The Wall Street Journal)
frequently report an announcment the day following its occurrence, day D - 1
is considered as well as day D. In cases of changes in realizable values (table 4,
panel A), D - 1 is associated with more negative mean prediction errors
( - 4 . 6 % ) than is the day of discovery (-2.2%). The day of discovery of a
fraud, mistake, or omitted transaction is associated with the most negative
mean daily prediction error in the tables ( - 5 . 1 % in table 4, panel B), but the
sum of the two daily mean prediction errors is more negative for changes in
realizable values (-6.8% in panel A to -6.0% in panel B). Although changes
in realizable values have been distinguished from other misrepresentations by
the courts, they do not appear to have any less of an effect on security prices
when they are discovered than do other misrepresentations. This conclusion is
reinforced by the fact that 13 of 18 changes in realizable values were associated
with negative prediction errors on the day of discovery, while this is true for
only 10 of 24 announcements of fraud, mistake, or omitted transactions.
The mean cumulative prediction errors ( M C P E ) reported in the right-hand
columns of the table 3 panels do indicate that changes in realizable values are
associated with greater prediscovery stock price declines than are fraud,
mistakes, and omitted transactions. On average, 95% (53.4% out of 56.1%) of
the negative returns associated with the damages claimed by plaintiffs alleging
that they were misled by a change in realizable values occurred prior to the
month the change was made public. Only about half (17.6% out of 34%), on
average, of the damages claimed for fraud, mistake and omitted transactions
relate to negative returns earned prior to the month of discovery. 16 of 17 cases

21Verrecchia (1980) theorized that the degree of precision of information, which he defined as
the inverse of the variance, is positively related to the concentration in time of capital market
reaction. Holthausen and Verrecchia (1983) have suggested that the precision of information
relative to previously available information is positively related to security price variability when
the information is announced.
198 R . L Kellogg, Class action lawsuits

Table 3
Mean monthly prediction errors (MPEo) for the last 25 months of damage computation periods
with mean cumulative prediction errors M C P E o classified by type of information announced at
discovery (D).
Panel A : Changes in realizable values ( T~'pe 4 information )
Month
( D = discovery
month) M P E ( + , - )a Z M C P E ( + , - )~ Z

D - 24 -0.019 3,3) 0.50 -0.272 (1,5) -2.01


D - 23 - 0.061 2,4) -1.82 -0.312 (1,5) -2.58
D - 22 - 0.021 1,6) - 0.62 - 0.277 (2,5) - 2.28
D - 21 - 0.012 4.4) 0.03 - 0.247 (3,5) - 1.77
D - 20 - 0.036 2,6) -1.55 -0.274 (1,7) -2.32
D- 19 -0.077 2,6) -3.19 -0.328 (1,7) -3.48
D - 18 -0.016 2,6) -0.29 -0.331 (2,6) -3.44
D - 17 - 0.064 2,6) -2.98 -0.379 (1,7) -4.32
D- 16 -0.040 2,6) -0.69 -0.394 (2.6) -4.37
D- 15 -0.109 2,7) -5.48 -0.405 (2,7) -5.39
D- 14 -0.060 3,7) -2.74 -0.392 (2,8) -6.12
D- 13 0.020 5,6) 1.71 -0.367 (2,9) -5.25
D - 12 - 0,064 6,6) -3.01 -0.358 (3,9) -6.22
D - 11 - 0.054 2,10) - 1.14 -0.361 (3,9) -5.51
D - 10 0.001 5,8) -0.66 -0.344 (4,9) -5.31
D - 9 - 0.052 6,9) - 2.31 - 0.308 (4,11) - 5.65
D- 8 -0.019 5,10) -2.05 -0.351 (5,10) -6.46
D - 7 - 0.069 4,11) - 2.52 - 0.385 (3,12) - 6.62
D - 6 - 0.050 6,9) - 1.77 -0.402 (4,11) -6.74
D - 5 - 0.085 2,13) - 3.84 - 0.440 (1,14) - 7.58
D - 4 -0.055 (6,10) - 3.33 - 0.426 (2,14) - 7.38
D - 3 - 0.073 (3,13) - 3.95 - 0.469 (2,14) - 8.55
D - 2 0.003 (6.1]) 0.38 - 0.481 (1,16) - 8.18
D - 1 - 0.091 (3,14) - 2.83 -0.534 (1.16) - 8.88
D - 0.048 (8,9) -1.36 -0.561 (1,16) -9.22

Average market model coefficients: .4 = 0.009 B = 1.20

Panel B: Fraud. mistake, and omitted transactions (Types 1, 2, and 3 information)


D - 24 - 0.047 (2, 8) - 2.06 0.059 (2, 8) - 3.53
D- 23 -0.033 (4,6) 0.28 -0.053 (2,8) -3.31
D - 22 0.059 (5, 5) 3.25 0.050 (2,8) - 2.70
D - 21 -0.034 (3,7) -0.70 -0.018 (2,8) -2.85
D- 20 -0.084 (3,8) -2.17 0.038 (2,9) -3.31
D - 19 -0.009 (6,6) 0.62 0.101 (2,10) -3.05
D - 18 -0.031 (5,7) -1.98 0.100 (2,10) -3.61
D - 17 -0.045 (3,11) - 1.22 0.215 (3,11) -3.93
D - 16 -0.001 (6,8) -0.16 0.246 (2,12) -3.82
D-15 -0.018 (6,10) -1.96 0.190 (ill) -3.10
D - 14 -0.047 (2,14) -1.54 0.031 (4,12) -3.39
D - 13 -0.020 (8,9) -0.56 -0.050 (3,14) -3.49
D-12 -0.033 (7,10) -1.19 -0.138 (3,14) -3.46
D - II -0.017 (9,9) -0.17 -0.125 (4,14) -4.06
D-10 0.018 (12,6) 1.32 -0.103 (5,13) -3.22
D-9 -0.000 (I0,8) -0.35 -0.010 (5,13) -3.16
D- 8 -0.016 (4,14) -1.36 -0.084 (5,13) -3.56
D-7 -0.032 (6,15) -2.38 -0.126 (5,16) -3.25
D- 6 -0.065 (8,13) -2.22 -0.126 (6,15) =3.80
D- 5 -0.059 (8,15) -3.06 -0.149 (4,19) -5.32
D- 4 0.052 ( 5 , 1 9 ) 3.13 -0.112 (5,19) -0.91
D - 3 - 0.026 ( 1 0 , 1 4 ) - 2.33 - 0.116 (5,19) - 2.01
D- 2 -0.055 (7.19) -2.36 -0.147 (6,20) -2.97
D - 1 - 0.046 ( 1 0 . 1 6 ) - 2.16 - 0.176 (6, 20) - 3.91
D -0.108 (5.21) -5.67 -0.340 (5,21) -5.53
Average market model coefficients: A = 0.001 B = 1.52

aSee footnote t o t a b l e 1.
Table 4
Mean daily prediction errors (MPE) and mean cumulative prediction errors (MCPE) classified
by type of information announced at discovery (D).
Day M P E ( + , - )a Z M C P E ( + . - )~ Z

Panel A : Changes in realizable oalues ( Type 4 information )


D - 20 0.007 (13,7) 1.11 0.007 (13,7) 1.11
D - 19 - 0.006 (11,9) -0.49 0.000 (12,8) 0.43
D - 18 0.017 (9,11) 1.37 0.017 (12,8) 1.15
D - 17 0.017 (10,10) 2.74 0.035 (11,9) 2.36
D - 16 0.004 (11,9) 0.45 0.039 (11,9) 2.32
D - 15 0.008 (11,9) 0.95" 0.045 (14,6) 2.50
D - 14 -0.002 9,11) 0.07 0.044 (12,8) 2.34
D - 13 -0.009 7,13) - 1.32 0.036 (8,12) 1.72
D - 12 0.002 11,9) 0.47 0.037 (12,8) 1.78
D - 11 - 0.011 5,14) - 1.76 0.025 (10,10) 1.14
D - 10 - 0.001 10,10) -0.40 0.023 (13,7) 0.96
D - 9 - 0.008 3,17) -1.51 0.017 (12,8) 0.49
D- 8 0.006 12,8) 1.31 0.021 (11,9) 0.83
D - 7 - 0.007 7,13) -1.18 0.016 (9,11) 0.48
D - 6 - 0.005 5,15) - 0.47 0.010 (10,10) 0.35
D- 5 -0.006 8.12) -0.66 0.004 (10,10) 0.17
D - 4 0.025 13,7) 3.24 0.030 (10, 10) 0.95
D - 3 - 0.003 (7,13) - 0.82 0.027 (10,10) 0.73
D - 2 - 0.007 (6,14) -0.86 0.022 (10,10) 0.51
D - 1 - 0 . 0 4 6 (10,9) -6.16 -0.043 (6,13) -1.31
D - 0.022 (5,13) -2.93 -0.076 (5,13) -1.85
D + 1 0.016 (7,10) 1.72 -0.063 (5,12) -1.33
D + 2 0.003 (10,7) 0.11 -0.060 (5,12) -1.28
D + 3 - 0 . 0 2 1 (8,10) - 2.66 - 0.081 (6,12) - 1.86
D + 4 0.030 (14,4) 3.69 -0.060 (6,12) -1.08
D + 5 0.002 (10,8) 0.48 -0.059 (7,11) -0.96
D+6 -0.006 (9,9) -0.95 -0.064 (7,11) -1.13
D + 7 0.015 (10,8) 1.94 -0.049 (6,12) -0.74
D + 8 0.003 (10,8) 0.88 -0.047 (8,10) -0.57
D + 9 0.010 (11,7) 1.46 -0.039 (7,11) -0.29
D + 10 0.008 (9,9) 1.56 -0.031 (8,10) -0.00
A v e r a g e m a r k e t model coefficients: A=-0.002 B= -1.26

Panel B: Fraud, mistake, and omitted transactions ( 7~vpes I, 2, and 3 information )


D - 20 - 0 . 0 1 6 (10,18) -2.67 -0.016 (10,18) -2.67
D - 19 0.012 (17,11) 1.51 -0.004 (12,16) -0.83
D - 18 0.011 (13,15) 0.50 0.007 (16,12) -0.39
D-17 - 0 . 0 2 1 (8,20) -2.78 -0.015 (9,19) - 1.72
D - 16 0.000 (16,12) 0.40 -0.017 (12,16) - 1.36
D - 15 0.015 (17,11) 1.86 -0.002 13,15) - 0.48
D - 14 0.002 (15,13) 0.37 0.001 11,17) -0.31
D-13 -0.008 (7,2t) -1.84 -0.007 13,15) - 0.94
D - 12 - 0 . 0 0 2 (13,15) -0.26 - 0.010 15,13) - 0.97
D - II 0.005 (17, II) I.I0 -0.007 14,14) -0.57
D - 10 0.004 (16,12) 1.38 -0.002 14,14) -0.13
D - 9 - 0 . 0 0 3 (15,13) -0.03 -0.006 14,14) -0.13
D - 8 0.006 (17,11) 0.68 - 0.000 12,16) 0.06
D - 7 0.009 (13,15) 0.40 0.007 15,13) 0.16
D - 6 - 0 . 0 0 1 (15,13) 0.21 -0.007 15,13) 0.21
D-5 - 0 . 0 1 7 (9,19) -1.85 -0.011 13,15) - 0.26
D - 4 0.017 (15,13) 1.59 0.008 14,14) 0.13
D- 3 - 0 . 0 0 0 (14,14) -0.60 0.007 12,16) -0.01
D - 2 - 0 . 0 1 4 (10,18) - 1.81 -0.008 12,16) - 0.42
D- 1 - 0 . 0 0 9 (12,15) -1.62 -0.025 12,15) -0.94
D - 0.051 (10,14) - 8.55 -0.075 8,16) -2.83
D + 1 0.003 (11,8) 0.83 -0.084 (5,14) -2.10
D + 2 - 0 . 0 1 3 (7,13) -4.41 - 0 . 1 0 1 (4,16) - 3.22
D + 3 0.004 (8,12) 1.24 - 0.101 (5,15) -2.90
D+4 - 0 . 0 2 9 (6,13) -3.30 - 0 . 1 1 6 (4,15) -3.10
D + 5 0.005 (11,9) 0.01 - 0.100 (3,17) - 2.77
D + 6 0.007 (8,12) 0.08 -0.095 (4,16) -2.71
D + 7 - 0 . 0 5 3 (10,11) -4.09 - 0 . 1 4 4 (5,16) -3.25
D + 8 0.003 (9,11) -0.29 - 0 . 0 9 8 (5,15) - 2.74
D+9 - 0 . 0 0 7 (10,9) -1.23 - 0 . 1 1 0 (5,14) -3.04
C + 10 - 0 . 0 0 8 (9,11) -2.71 - 0 . 1 2 1 (7,13) - 3.93
A v e r a g e m a r k e t model coefficients: A ~ - 0.001 B = 1.59

aSee footnote to table 1.


200 R.L Keilog~ Class action lawsuits

involving Type 4 information are characterized by negative cumulative returns


prior to the month of misrepresentation discovery. 20 of the 26 cases involving
fraud, mistake, or omitted transactions exhibit negative prediscovery cumula-
tive returns.
A T-test is performed on the difference between the percentage of total
negative returns occurring prior to the month of discovery (%) in Type 4 cases
and that observed in cases involving fraud, mistake, and omitted transaction
where

% = MCPE o_ x/MCPED.

The test is conducted using only the 16 cases involving realizable values and
the 21 cases involving the other three types of discovery information for which
cumulative prediction errors were negative at the end of the month of dis-
covery. For these firms the mean percentages are 92% and 83%, respectively.
The T-statistic is 1.33 which does not permit rejection of the null hypothesis of
no difference between the mean percentages with less than a 5% chance of
error (one-tailed test with 35 degrees of freedom). The null hypothesis can be
rejected at the 10% level. The chi-square statistic for the association of the sign
( + or - ) of prediscovery returns with type of misrepresentation is not
significantly different from zero (X 2= 1.15).

5. Summary and discussion


Discoveries of what are alleged in lawsuits by common stock purchasers to
be misrepresentations in financial statements are associated with significant
negative risk adjusted common stock returns. Lawsuit plaintiffs would argue
that this is evidence that financial statement misrepresentations misled inves-
tors. Negative returns are also observed months before misrepresentations are
discovered. Plaintiffs would argue that corresponding stock price declines are
effects of selective leakages of information that prior financial statements
contained misrepresentations and that associated portfolio losses are properly
included in damage computations. An alternative hypothesis is that lawsuits
are more likely to occur when negative returns have preceded financial state-
ment revisions; in such cases plaintiffs have opportunities to claim damages for
prediscovery stock price declines.
Certain accounting conventions that require periodic asset valuations have
produced revisions in financial statements from year to year that have been
cited by lawsuit plaintiffs as evidence of misrepresentation. There is some
evidence that announcements of these revisions are associated with different
patterns of security returns than are other kinds of announcements that
contribute to lawsuit occurrence. On average, information that asset realizable
values are being revised is associated with greater abnormal returns prior to
R.L. Kellog~ Class action lawsuits 201

announcement and smaller abnormal returns in announcement months than


are announcements that fraud, mistake, or failure to separately note specific
transactions affected prior financial statements.
In accounting buyers' suits courts have to decide whether a particular
financial statement element affects stock prices. There is little in the accounting
literature that can inform them in making these decisions. Empirical account-
ing research has concentrated on the associations of earnings changes and
changes in accounting methods with security prices. 22 I am aware of no
published research that compares the returns series associated with earnings
changes resulting from different events.
The lack of empirical evidence on this issue is understandable in view of the
lack of theory that relates accounting numbers to asset prices. In distinguishing
changes in realizable values from other kinds of information that was an-
nounced at misrepresentation discoveries, I relied on a similar distinction that
had been made by previous observers of and participants in the legal treatment
of accounting issues. The distinction is ad hoc. Only recently have attempts
been made to identify factors that affect the relationship between accounting
information and security prices. 23 Development of theory in this area and
research that relates specific items of financial statements to security prices is
critical to rationalizing the legal process that is applied to financial reporting
activities.

Appendix
The firm name, four-digit SIC code, market values of outstanding common
shares at beginning and end of damage computation period (DCP), number of
months in damage computation period, and type of information announced at
discovery (as discussed in section 4), for each of the 56 firms for which returns
data are reported, are listed below. Based on the available sources, classifica-
tion by information type was not straightforward. For example, the Kleinerts
lawsuit involved 'inventory problems in the Danoca Division'. This was
classified as Type 2 (mistake) since there is no explicit mention of intentional
or illegal acts. Distinguishing Type 1 from Type 2 cases does not affect the
returns data that are reported. But correctly distinguishing Type 4 (changes in
asset realizable values) is important to any interpretation of the results in
section 4. Difficult Type 4 classification decisions involved Alaska Interstate
('cost overruns and inventory adjustments'), Lockheed (contract bid was 'so
low it insured a loss') and Mortgage Growth Investors (allegations of con-
spiracy involving advisory fees in addition to inadequate reserves for risky

22See Kaplan (1978) for a review of this research.


23See Copeland (1976), Goldman and Sosin (1979), Verrecchia (1980), Hol~ausen and
Verrecchia (1983).
202 R. L Kellogg Class action lawsuits

Table A.1

Market value
(in $ millions)

Start End Months


SIC of of in Info.
Firm name code DCP ~ DCP DCP type ~ '

Alaska Interstate 3536 98.6 53.2 11 4


American Fletcher Mortg. 6162 30.2 5.7 27 4
American Realty Trust 6531 18.0 9.5 3 4
Ampex 3679 243.6 119.6 21 2
BT Mortgage 6162 61.8 9.0 13 4
Bankers Trust 6025 526.3 314.6 22 4
Bermec 7513 101.6 49.7 14 1
Botany Inds. 2311 18.4 5.9 28 3
C N A Financial 6321 577.0 393.0 46 4
Capital Mortg. Inv. 6162 31.6 7.3 10 1
Cenco 3811 501.2 65.9 38 1
Chris Craft 3732 42.7 20.0 40 4
C.I. Mortgage 6799 115.0 19.2 10 4
Combustion Eng. 3499 587.6 471.5 16 3
Continental II1. Rlty. 6799 50.5 17.5 51 4
Continental Mortg. 6799 194.7 46.8 16 4
DPF 7394 134.0 20.2 14 4
Ecological Science 3430 67.4 46.5 10 4
FAS Int'l. 8220 136.9 16.8 30 4
Faberge 2844 96.9 78.1 3 2
First Mtg. Invs. 6799 32.5 8.2 62 1
First Pennsylvania 6025 529.8 198.0 25 1
GAC 6145 191.4 30.6 49 1
GSC Enterprises 5912 8.2 3.8 80 3
Genesco 2341 282.7 175.5 82 1
Horizon 6552 135.0 15.1 61 1
ICN Pharmaceutical 2834 192.0 82.9 12 4
IDS Realty 6799 36.2 13.0 9 4
International Controls 3662 38.0 11.9 8 l
Interstate Stores 5311 5.5 2.7 6 3
Kleinerts 3069 7.1 19.4 18 2
K oracorp 2335 20.4 10.9 16 2
Lockheed 3721 653.1 715.0 5 4
Marine Midland 6025 298.4 166.7 25 4
Mattel 3944 280.0 37.6 77 1
Memorex 3679 305.5 150.2 12 1
Mohawk Data 3573 107.25 224.3 8 2
Mortgage Growth Inv. 6799 20.15 10.9 16 4
Noel Inds. 2328 5.3 2.8 5 4
Occidental Petrol 1311 1739.5 1130.9 20 2
Paterson Parchment Paper 2621 20.1 4.9 33 2
Polaroid 3861 2113.8 797.8 3 3
Realty Equities 6519 42.1 0.7 14 1
Richton Int'l. 5137 20.6 27.2 3 4
Rockwood Nat'l. 7379 19.6 3.7 46 1
g u s c o Industries 3442 8.5 14.8 56 3
Sanitas Service 7349 47.9 36.8 8 3
Signal Cos. 3711 622.6 303.9 23 4
Stratton Group 5331 6.0 2.7 18 2
Susquehanna 3079 206.5 26.4 64 4
T.I. Corp. 6361 171.4 81.7 37 1
Teleprompter 4833 402.0 246.0 6 1
U.S. Realty 6531 47.1 8.6 62 4
Valley Metalurgical 3399 58.6 32.0 3 2
Western Union 4821 718.8 133.9 26 4
Whittaker 1742 261.9 194.0 5 1

a D C P refers to the Damage C o m p u t a t i o n Period.


b Information types: 1. fraud, 2. mistake, 3. omitted transactions, 4. changes in realizable values.
R.L, Kellogg Class action lawsuits 203

loans). Changes in asset realizable values appeared to be the predominant type


of information announced at discovery in these cases. Real estate investment
trusts (SIC codes 6531, 6162, and 6799) were associated with both Type 1 and
Type 4 discoveries. In each case, plaintiffs alleged inadequate provision for
loan losses. Cases classified as Type 1 involved allegations of activities under-
taken to conceal specific uncollectible loans.

References
Anderson, J., 1977, The potential impact of knowledge of market efficiency on the legal liability of
auditors, The Accounting Review 52, 417-426.
Antle, R., 1982, The auditor as an economic agent, Journal of Accounting Research, Autumn,
503-527.
Ball, R., 1971, Changes in accounting techniques and stock prices, Empirical research in account-
ing: Selected studies 1972, Supplement to Journal of Accounting Research, 1-38.
Ball, R. and P. Brown, 1968, An empirical investigation of accounting income numbers, Journal of
Accounting Research, Autumn, 159-178.
Brown, P., 1970, The impact of the annual net profit on the stock market, The Australian
Accountant, July, 277-282.
Carmichael, D., 1976, Litigation and the auditor's responsibility, Journal of Accountancy 142,
July, 75-76.
Chambers, R., 1974, Stock market prices and accounting research, Abacus, June, 39-74.
Copeland, T., 1976, A model of asset trading under the assumption of sequential information
arrival, Journal of Finance, Sept., 1149-1168.
Fischel, D., 1982, Use of modern finance theory in securities fraud cases involving actively traded
securities, Business Lawyer 38, 1-20.
Foster, G., 1975, Accounting earnings and stock prices of insurance companies, The Accounting
Review, Oct., 686-698.
Foster, G., 1980, Accounting policy decisions and capital market research, Journal of Accounting
and Economics, March, 29-62.
Goldman, M. and H. Sosin, 1979, Information dissemination, market efficiency and the frequency
of transactions, Journal of Financial Economics 7, 29-61.
Holthansen, R. and R. Verrecchia, 1983, The change in price resulting from a sequence of
information releases, Working paper, March (University of Chicago, Chicago, IL).
Jaenicke, H., 1977, The effect of litigation on independent auditors (Commission on Auditors'
Responsibilities, New York).
Kaplan, R., 1978, The information content of financial accounting numbers: A survey of empirical
evidence, in: A. Abdel-Khalik and T. Keller, eds., The impact of accounting research on
practice and disclosure (Duke University, Durham, NC) 134-173.
Kripke, H., 1971, Rule 10b-5, Liability and 'material facts', NYU Law Review 46, 1061-1076.
Pattell, J., 1976, Corporate forecasts of earnings per share and stock price behavior: Empirical
tests, Journal of Accounting Research 14, 246-276.
Scholes, M. and J. Williams, 1977, Estimating betas from nonsynchronous data, Journal of
Financial Economics, Dec., 309-328.
Simunic, D., 1980, The pricing of audit services: Theory and evidence, Journal of Accounting
Research 18, 161-190.
St.Pierre, K. and J. Anderson, 1984, An analysis of the factors associated with lawsuits against
public accountants, Accounting Review, April, 242-263.
Verrecchia, R., 1980, The rapidity of price adjustments to information, Journal of Accounting and
Economics, March, 63-92.
Wallace, 1980, The economic role of the audit in free and regulated markets (Touche Ross & Co.,
New York).

J.A.E.-- C
204 R . L Kellogg, Class action lawsuits

C_AIse$
Blue Chip Stamps v. Manor Drug Stores, 421 U.S. 723 (1975).
Chasins v. Smith, Barney, & Co., 306 F. Sup. 177 (SDNY, 1969).
Ernst & Ernst v. Hochfelder, 425 U.S. 185 (1976).
Esplin v. Hirschi, 402 F. 2d 94 (10th Cir. 1968).
Harris v. American Investment Co., 523 F. 2d 220 (8th Cir. 1975)
Lewis v. Black, CCH Fed. Sec. L. Rep. P95, 312 (EDNY, 1976).
McLean v. Alexander, 599 F. 2d 1190 (ed Cir. 1979).
Marx v. Computer Sciences Corp., 507 F. 2d 485 (9th Cir. 1974).
Polin v. Conductron Corp., 552 F. 2d 797 (Sth Cir. 1977).
Sanders v. John Nuveen & Co., 554 F. 2d 790 (7th Cir. 1977).
U.S.v. Benjamin, 328 F. 2d 854 (2d Cir. 1964).

You might also like