Proceedings of Applied International Business Conference 2008
EARNINGS MANAGEMENT AND INFORMATION CONTENT OF TIMING THE ANNUAL
AND INTERIM EARNINGS RELEASES: THE CASE OF MALAYSIA
Shamsher Mohamad and Cheng Fan Fah ψ
Universiti Putra Malaysia, Malaysia
___________________________________________________________________________________
Abstract
This paper examines how earnings management affects the share markets’ reaction to earnings
announcements by timing the announcement dates and the information content of timing the
announcement beyond the actual earnings announced. The annual and interim earnings response
coefficients of listed firms in Bursa Malaysia provide evidence on the relationship between the timing
of earnings announcement and the direction and magnitude of annual and interim earnings for more
than 4912 firm-years. The results confirm that firms time their earnings announcement based on the
direction and magnitudes of the unexpected earnings. Firms announce earnings early for positive
unexpected earnings, and delay the announcement for negative unexpected earnings. The market reacts
to the timing of the announcements differently. The timing of annual earnings received more vigorous
responses than interim earnings announcements. These findings are relevant and useful to judge
company performance by looking at the timing of announcement the company earnings.
___________________________________________________________________________________
Keywords: Annual and interim report release; Earnings management; Directional and magnitude;
Timing; Earnings response coefficients.
JEL Classification Codes: G14; G24; G3.
1. Introduction
Price effect of information in developed capital markets has been extensively researched and has
documented that price changes are affected by information and that accounting earnings is a subset of
information affecting share prices. Recently, evidence from the Asia Pacific region (Cheng et al 2001,
Arif et al 1997) have documented a robust price-to-earnings relationship but of lesser magnitude than
that documented in developed markets.
One of the desired qualitative characteristics of financial reporting is timeliness. In fact, managers place
much importance on the timing of earnings announcement and adjust their earnings announcement
dates (Chen and Mohan (1994). The theory of earnings management and conservation, that is, whether
bad news in financial statements is captured faster than good news is well documented in the literature.
Basu (1997) reported that firms delay the announcement of unexpected loses (“bad news”) or speed up
the announcement of unexpected gain (“good news”). These findings are consistent with the
‘Conservatism’ in the accounting discipline that dictates pessimistic attitude when choosing accounting
techniques for financial reporting.
Whilst annual and interim reports are important source of information to stakeholders, their relevance
depends on the timely release of these information. To further substantiate the evidence on the issue of
earnings management and timing the release of information through annual and interim earnings
reports in an emerging market, this paper address this issue using a sample of annual and interim
earnings announcements of firms listed on Bursa Malaysia, and its impact on firm’s value.
2. Literature review
Sinclair and Young (1990) examined the association between the timeliness of half-yearly report for
Australian firms and the abnormal stock price behavior around the time of the announcement. They
conclude that ‘good’ news is released earlier than ‘bad’ news.
Haw, Qi and Wu (2000) analyzed the timeliness of annual report releases and market reaction to
earnings announcement in China. They reported that firms with good news release their annual reports
ψ
Corresponding author. Cheng Fan Fah. Department of Economics, Faculty Economics and
Management, Universiti Putra Malaysia, 43400 UPM, Serdang, Selangor, Malaysia. Corresponding
author Email: [email protected]
Proceedings of Applied International Business Conference 2008
earlier than firms experiencing losses (bad news). They also observe a significant price reaction to
annual earnings announcements for both early (good news) and late (bad news) reporting firms. The
findings are consistent with Chambers and Penman (1984) and Bagley and Fisher’s (1998) findings
that firms unexpectedly accelerate the release of good news and delay the disclosure of bad news
relative to their previous reporting pattern.
Chen, Cheng and Gao (2005) find that firms that are willing to make early announcements, tends to
surprise the market, as indicated by the higher volume and price reactions, compared to lower volume
and price reactions for later announcements. Their results indicate that an information asymmetry exists
between early and late earnings announcements in China.
There are many studies on the information content of other accounting numbers beyond accounting
earnings. Jenning’s (1990) review of two previous studies (Rayburn 1986; Bowen et al. 1987) provides
consistent and strong evidence that both cash flow and accrual components add to the in formativeness
of income. However, there is only weak and inconsistent evidence that accrual components of income
are valued differently from cash flow components by investors.
Ali (1994) estimates the linear model of annual stock returns with changes in earnings, working capital
from operation, and cash flows. The results suggest that earnings have incremental information content
beyond working capital and cash flows, and working capital has incremental information content
beyond earnings and cash flows. But, the cash flows have no incremental information relative to
earnings and working capital.
Ali and Pope (1995) re-examined another incremental information content study of earnings, fund
flows and cash flows by using non-linear model for UK data, which include current level of earnings
together with the change in earnings as complementary proxies for the unexpected components of
earnings. The results show that earnings have higher relative information content than both fund flows
and cash flows. Other studies by Collins and Clubb (1995) and Cotter (1996) find mixed results. In the
US, there is ample evidence on the information content of interim financial reports (Hagerman,
Zmijewski and Shah 1984, Swaminathan and Weintrop 1991).
Opong (1995) examines the issue of whether interim financial reports in the UK contain value relevant
information and concludes that interim financial reports contain material information on the day the
reports are released. Schadewitz (1996) investigates the association between interim earnings and
security prices in Finland and reports that interim earnings explained unexpected interim returns.
Interim reports are not audited or are prepared long time before an audit is done, therefore the contents
might be too tentative and subjective without a potential audit opinion hence should be less important
than annual earnings.
Lev’s (1989) review of relevant theories and evidence on the usefulness of earnings research over the
1980 and 1988 period suggests that though investors do use earnings reports to gauge the financial
health of a firm, the extent of earnings usefulness is rather limited.
Findings on the returns-to-earnings relationship is well documented in the literature and characterized
by continuous attempts to further validate this relationship through refinements, such as controlling for
revenue (Swaminathan and Weintrop, 1991), firm size (Chaney and Jeter, 1993), leverage effect
(Dhaliwal, Lee and Farger, 1991) and auditor choice (Shamsher and Cheng, 2001).However, the
evidence from these studies is inconclusive.
3. Objectives
The major objective in this paper is to ascertain that firms unexpectedly accelerate the release of good
news and delay the disclosure of bad news relative to their previous reporting pattern. The abnormal
returns will also vary according to the dates of the announcements. The second objective is to ascertain
whether changes in announcement dates affect share prices in response to the information content in
the announcements..
4. Data and methodology
Sample Selection
The earnings announcements dates are collected from the Bursa Malaysia’s (formerly known as the
Kuala Lumpur Stock Exchange) Daily Dairy for a random sample of 160 firms over a 10-year period
1229
Proceedings of Applied International Business Conference 2008
(Jan 1988 to July 1997). The total sample size is 4914 earning announcements, comprising of 2482
annual earnings announcements and 2431 interim earnings announcements. After screening and removal
of non-earnings-related disclosures the sample consists of 430 annual earnings announcements and 254
interim earnings announcements.
Methodology
The impact of timing of earnings announcements and earnings changes on stock price is evaluated using
the standard event study methodology. The abnormal returns are estimated by taking the difference in
the realized and expected returns. The expected return is estimated using market model as shown in
equation 1.
Abnormal Returns:
ARit = Rit - [αi + βi Rmt]
(1)
Where, AR is the abnormal returns
with Rit = Ln (Pit/Pi t-1) and Rmt = Ln (It / It-1),
where, in addition to terms already defined, Ln is natural logarithm and I refers to market's composite
index.
The market parameters αi and βi are estimated over trading periods, –60 months to –3 month relative to
the announcement month. The returns were adjusted for thin-trading bias using Fowler-Rorke’s method
(1983). The resulting risk-adjusted abnormal returns of each observation at any time over the test
window is added and averaged across all the observations to obtain the average abnormal return (AARt).
The accumulation is done over a price reaction window of –50 to +1(CAR 50).
Next the average abnormal returns over t = 1, …., T is cumulated as follows:
CAR
where,
=
T
∑ AARi t
t=1
CAR is the cumulative average abnormal return
AAR it is the average abnormal return for firm i at time t.
(2)
Analysis of Unexpected Accounting Earnings
Unexpected earnings are computed using the naive expectation model, which assumes that the next
period's expectation is simply the current period's earnings. This is also consistent with the design of the
study to investigate the contemporaneous effect of price change at a point in time.
Raw unexpected earnings (RUEs) are computed using the naive model:
RUEit
Eit
= Eit - Ei(t-1)
= Earnings for firm i at time t.
(3)
The unit normal variables are estimated as follows:
where
SUEi
σ(UEi)
= RawUEi/σ(UEi)
is standard deviation of UE
(4)
This transformation, which mitigates the effect of changing variance or heteroscedasticity on the
variables, yields unexpected value of earnings variable adjusted for volatility differences, σ(UEi).
Analysis of Shift in the Earnings Announcement dates
In this paper two dates of announcement are used; the first date is the day of announcement from the
year-end dates (DYE) and second is the day of announcement from the mean announcement dates,
which is the unexpected shift in earnings announcement dates (UAD).
Unexpected shift in earnings announcement dates are computed using current period's earnings
announcement dates less the previous period average announcement dates. The values will be either
negative or positive, where negative means a shift in earlier announcement and positive means a delay in
announcement. This is also consistent with the research design to study the contemporaneous effect of
price change at a point in time.
1230
Proceedings of Applied International Business Conference 2008
Unexpected earnings announcement dates (UADit) are computed using as follow:
UADit
where
= ADit - MADi(t-1)
(5)
UADi is Unexpected Announcement dates, in days
AD
is Announcement dates, in days
MAD is Average announcement dates, in days.
Sample Grouping
Grouping of samples has an advantage of reducing the errors-in-variables problem: (Beaver et al.,
1979, 1980, 1987; Ariff et al., 1997). The portfolios were formed from the sampled firms first by
ranking all stocks according to the magnitude of unexpected announcement dates (UAD). To form 4
quartiles, the twenty five percent of the stocks with the highest rank on the UAD ranking are placed in
the first quartile. The next highest twenty five percent in the second quartile, and so forth until the
fourth quartile contains the last twenty five percent of the observations with the lowest unexpected
announcement dates (UAD).
Then an independent sample test is conducted on the difference of mean for the first and the fourth
quartile on the days of announcement from financial year-end, the days of announcement from the
previous mean announcement dates, the raw unexpected earnings, the standardized unexpected
earnings, the CAR(50) and CAR (1).
The relationship between the unexpected earnings and cumulative abnormal returns is examined using
equation (6).
CARit = a + b*SUEit + eit
(6)
where,
CARit : is some measure of risk-adjusted return for security i cumulated over period t,
SUEit : is a measure of standardized unexpected earnings, and
: is a random disturbance term assumed to be normally distributed.
eit
The slope coefficient of the regression, b, is called the earnings response coefficient (ERC).
Timing in Earnings announcement dates
The shift in earnings announcement dates is assumed to be synonymous with the firms timing their
earnings announcement based on the magnitudes of the unexpected earnings. Therefore shift in
earnings announcement dates variable is included into the regression to test the effect of timing of
earnings announcement on ERC on listed firms in the Bursa Malaysia. Two dummy variables (D) are
used to signify the early and late earnings announcements in the regression.
CAR i = a1 + a2 SUE i + a3 De i + a4 Dl i +ε i
(7)
where, CARi : Cumulative Abnormal returns over a specified window,
SUEi : Standardised Unexpected Earnings,
De
: Dummy variable taking a value of :
= 1 for firms that announce earnings early (1st quartile firms), and
= 0 for others.
Dl
: Dummy variable taking a value of
= 1 for firms that announce earnings late (4th quartile firms), and
= 0 for others.
A significant coefficient for early announcement dates a3, or late announcement dates a4 would support
the notion that timing has information content beyond unexpected earnings. A positive value is
expected for a3 as firms that earn positive unexpected earnings are expected to make early
announcements. A negative value for a4 is expected as firms with negative earnings are expected to
delay earnings announcements.
5. Findings
Annual earnings announcements
Panel A of Table 1 shows the descriptive statistics for the days of annual announcement from the
financial year-end, the unexpected earnings and the cumulative abnormal returns. The days of annual
announcement from the financial year-end have a mean of 89 days (approximately 3 months). The
1231
Proceedings of Applied International Business Conference 2008
standard deviation is 23 days. The minimum number of days for firms to make their annual
announcement of their year-ends results is 36 days, slightly more than a month. The maximum number
of days for firms to make the year-end results annual announcement is 161 days (approximately more
than 5 months) in this sample. The descriptive statistics also show that some firms have shifted the
annual announcement day earlier by 48 days and some have delayed their annual announcement dates
from previous period mean annual announcement dates by 69 days. The raw unexpected earnings have
a mean and standard deviation of 2.576 cents per share and 10.918 cents per share respectively. The
minimum raw unexpected earnings of firm is -40.7 cents per share (loss). The maximum raw
unexpected earning is 50.2 cents per share (gain). The 50 days risk adjusted cumulative abnormal
earnings has a mean and standard deviation of 0.4% and 11.2% respectively. The minimum riskadjusted cumulative abnormal return is –36.1% and the maximum risk-adjusted abnormal accumulative
abnormal return is 35% during earnings announcement.
Table 1: Descriptive statistics for days of announcement, unexpected earnings and cumulative
abnormal returns for full test sample (Panel A), 1st quartile (Panel B) and 4th quartile (Panel C)
Panel A: Full Test Sample
Interim Test Sample =254
Annual Test Sample: n = 430
Indices
Mean
Std. D.
Minimum Maximum Mean
Std. D.
Minimum
DYE
89
23
36
UAD
0
15
-48
RawUE
2.576
10.918
-40.700
SUE
0.291
0.903
-2.590
CAR50
0.004
0.112
-0.361
Panel B: 1st Quartile (Early Announcements)
Annual Test Sample: n = 104
Indices
Mean
Std. D.
Minimum
DYE
UAD
RawUE
SUE
CAR50
78
-16
3.816
0.412
0.034
14
9
8.194
0.857
0.115
36
-48
-20.800
-2.125
-0.236
Panel C: 4th Quartile (Late Announcements)
Annual Test Sample: n = 104
Indices
Mean
Std. D.
Minimum
DYE
113
UAD
19
RawUE
0.964
SUE
0.132
CAR50
-0.024
Indices: DYE
UAD
RawUE
SUE
CAR(50)
21
13
6.28
0.87
0.1889
161
69
50.200
3.106
0.350
67
0
1.23
0.21
0.0100
Maximum
Interim Test Sample =65
Mean
Std. D.
Minimum
Maximum
123
-6
49.500
2.812
0.339
52
-15
1.66
0.46
0.0313
94
-6
13.98
2.56
1.2643
Maximum
Interim Test Sample =65
Mean
Std. D.
Minimum
18
9
5.93
0.92
0.2125
21
-52
-26.00
-2.59
-1.0342
Maximum
21
-52
-26.00
-2.28
-0.3671
23
57
161
85
19
33
15
6
69
15
11
6
10.667
-36.900
43.000
0.59
7.32
-20.10
0.838
-2.590
2.383
0.28
0.71
-1.51
0.115
-0.306
0.276
0.0145
0.1654
-0.4979
= the number of days from announcement dates to Financial year-end dates
= the number of days between the announcement dates to the previous
mean announcement dates
= Raw Unexpected Earnings = this year earnings minus last year earnings
= Standardized Unexpected Earnings
= Cumulative Abnormal Returns from -50 to +1 days.
Panel B of Table 1 shows the descriptive statistics of the 1st and 4th quartile of the test sample. The 1st
quartile of the sample consists of 104 firms that announce annual earnings earlier by 6 to 48 days from
the previous average annual announcement dates (mean announcement dates) or these firms announce
their annual earnings from 36 to 123 days from their year-end dates. Their mean UAD is 16 days
earlier than their mean annual announcement dates or 78 days from the year-end dates. These firms
have raw unexpected earning of 3.816 cents per share respectively. The mean CAR50 is 3.4%.
1232
164
76
32.50
2.56
1.2643
Maximum
164
76
32.50
1.97
0.6042
Proceedings of Applied International Business Conference 2008
Panel C of Table 1 shows the descriptive statistics of the 4th quartile of the test sample. The 4th quartile
consists of 104 firms that have delayed the annual earnings announcement. The 4th quartile firms delay
their annual earnings announcements by 6 to 69 days from their previous average annual announcement
dates or these firms announce their annual earnings from 57 to 161 days from their year-end dates.
Their mean UAD is 19 days later than their mean annual announcement dates or 113 days from the
year-end dates. These firms have a mean raw unexpected earning of 0.964 cents per share. The mean
CAR50 is -2.44%
Interim earnings announcements
Panel A of Table 1 (columns 6 to 9), shows the descriptive statistics for the days of interim
announcement from the mid-financial year-end, the unexpected earnings and the cumulative abnormal
returns. The days of interim announcement from the mid-financial year-end have a mean of 67 days
(approximately 2 months). The standard deviation is 21 days. The minimum number of days for firms
to make their interim announcement of their mid-year-ends results is 21 days, slightly less than a
month. The maximum number of days for firms to make the mid-year-ends results interim
announcement is 164 days (approximately more than 5 months) in this sample. The descriptive
statistics also show that some firms have shifted the interim announcement day earlier by 52 days and
some have delayed their interim announcement dates from previous period mean interim announcement
dates by 76 days. The raw unexpected earnings have a mean and standard deviation of 1.23 cents per
share and 6.28 cents per share respectively. The minimum raw unexpected earning (RUE) of firm is 26 cents per share (loss). The maximum raw unexpected earning is 32.5 cents per share (gain). The 50
days risk adjusted cumulative abnormal earnings has a mean and standard deviation of 0.01% and
18.9% respectively. The minimum risk-adjusted cumulative abnormal return is –103% and the
maximum risk-adjusted abnormal accumulative abnormal return is 126% during interim earnings
announcement.
Panel B of Table 1 (columns 6 to 9), show the descriptive statistics of the 1st and 4th quartile of the test
sample. The 1st quartile of the sample consists of 104 firms that announce interim earnings earlier by 6
to 52 days from the previous average interim announcement dates (mean interim announcement dates)
or these firms announce their interim earnings from 21 to 94 days from their mid-year-end dates. Their
mean UAD is 16 days earlier than their mean annual announcement dates or 78 days from the midyear-end dates. These firms have raw unexpected earning of 1.66 cents per share. The mean CAR50 is
3.13%.
Panel C of Table 1 (columns 6 to 9) shows the descriptive statistics of the 4th quartile of the test sample.
The 4th quartile consists of 104 firms that have delayed the interim earnings announcement. The 4th
quartile firms delay their interim earnings announcements by 6 to 76 days from their previous average
interim announcement dates or these firms announce their interim earnings from 33 to 164 days from
their mid-year-end dates. Their mean UAD is 19 days later than their mean annual announcement dates
or 113 days from the year-end dates. These firms have a mean raw unexpected earning of 0.59 cents
per share. The mean CAR50 is –1.45%
Table 2: Frequency Distribution of Earnings Announcement based on Days after the Mean
Announcement Date (UAD): Annual (Panel A) and Interim (Panel B)
Panel A: Annual: n = 430
Days
<-30
-30 to -15
-15 to -5
-5 to +5
+5 to +15
+15 to +30
>+30
Number
Percent
10
2%
37
9%
76
18%
203
47%
63
15%
23
5%
19
4%
-15 to -5
40
16%
-5 to +5
116
46%
+5 to +15
47
19%
+15 to +30
22
9%
>+30
4
2%
Panel B: Interim: (n = 254)
Days
<-30
-30 to -15
Number
7
18
Percent
3%
7%
1233
Proceedings of Applied International Business Conference 2008
250
0 .5
0 .4 5
200
0 .4
150
0 .3
0 .2 5
100
0 .2
Percentage
Number
0 .3 5
0 .15
50
0 .1
0 .0 5
Annual
0
0
Days
<-3 0
-3 0 to -15
-15 to -5
-5 t o +5
+5 to +15
+15 t o +3 0
>+3 0
Interim
% Annual
Days
% Interim
Figure 1: Frequency distribution of annual and interim announcement test sample from mean
announcements dates (UAD)
Table 2 shows the frequency distribution of the earnings announcement date based on the day after the
financial year-ends (Panel A) and previous mean announcement dates (Panel B). Panel A shows that all
the frequency distribution of firms that announce their earnings early and later than their previous mean
announcement dates. Most of the firms have maintained their announcement dates within five days
from their previous mean announcement dates (47%). Some of the firms (18%) have announcements
earlier by 5 to 15 days and some (15%) have delayed their announcement by more than 5 to 15 days.
Panel B shows the frequency distribution of the interim earnings announcement from their mean
announcement dates. Most of the firms have maintained their announcement dates within five days
from their previous mean announcement dates (46%). Some of the firms (16%) have announcements
earlier by 5 to 15 days and some (19%) have delayed their announcement by more than 5 to 15 days.
The two distributions are close to each other. Figure 1 plotted the distribution and shows that they are
of similar shapes and close to normal.
Table 3: Independent samples test for 1st quartile (early announcement firms) and 4th
quartile (delay announcement firms) samples t-test for equality of means
Annual
Interim
Signficant
(2-tailed)
Mean
tIndices
Mean Difference t-values Signficant (2-tailed) p-values Difference values p-values
D Yr-end
-35
-13.087
0.000***
UAD
-35
-20.258 0.000***
RawUE
2.852
2.165
0.032*
SUE
0.280
2.381
0.018*
CAR50
0.057
3.605
0.000***
Note: Significant at p=0.05 (*), p=0.01 (**), and p=0.001 (***) levels.
For details of indices, please refer to Table 4.
-33
-33
1.075
0.179
0.017
10.37
17.41
0.938
1.280
0.519
0.000***
0.000***
0.350
0.203
0.605
Table 3 (columns 2 to 4) summarize the independent samples t-test for the 1st quartile (firms that
announce annual earnings early) and the 4th quartile (firms that announce annual earnings late). The
results show that the days from the annual announcement dates to the financial year-ends have a t-value
of -13.087, which is significant at 0.000 level. Similarly, the t-statistics for the days of annual
announcement from their previous mean annual announcement dates is –20.258, significant at 0.000
1234
Proceedings of Applied International Business Conference 2008
level. The raw unexpected earnings and the standardized unexpected earnings have t-statistics of 2.165
and 2.381 respective which are significant at 0.05 level. The 50-days risk-adjusted cumulative returns
(CAR50) have a t-statistic of 3.605, significant at 0.000 level. All the above results indicate that the
firms that announce annual earnings early are different from the firms that announce annual earnings
late in term of their unexpected earnings and the share price revaluation. The magnitudes are such that
the firms that announce annual earnings early have a higher raw unexpected earning (RUE) and larger
cumulative abnormal return for unexpected return (CAR50). This is consistent with the literature in
develop markets that firms time their earnings announcements dates based on the direction and
magnitudes of their earnings. The pricing behavior of Malaysian listed firms is similar to the firms in
other developed countries, that is the tendency of management to delay announcing bad news.
Table 3, (columns 5 to 7), summarize the independent samples t-test for the 1st quartile (firms that
announce interim earnings early) and the 4th quartile (firms that announce interim earnings late). The
results show that the days from the interim announcement dates to the mid-financial year-ends have a tvalue of –10.37, which is significant at 0.000 level. Similarly, the t-statistics for the days of interim
announcement from their previous mean interim announcement dates is –17.41, significant at 0.000
level. The raw unexpected earnings and the standardized unexpected earnings have t-statistics of 0.938
and 1.280 respective which are not significant at 0.05 level. The 50-days risk-adjusted cumulative
returns (CAR50) have a t-statistic of 0.519, not significant at 0.05 level. The above results indicate that
for the interim announcement dates, the firms that announce interim earnings early are not different
from the firms that announce interim earnings late in terms of their unexpected earnings and the share
price revaluation. The magnitudes are such that the firms that announce interim earnings early have a
same raw unexpected earning (RUE) and same cumulative abnormal return for unexpected return
(CAR50). This is inconsistent with the findings on annual earnings announcements dates based on the
direction and magnitudes.
Table 4:
Model
Independent
Variable
Regression Results For Returns-to-Earnings Relation For Firms announce
earnings early( 1st Quartile) and late (4th Quartile)
Regression Model: CAR i = a1 + a2 SUE i+ εii
Dependent Variable: Abnormal returns (AR),
Independent Variable: Standardized Unexpected Earnings (SUE)
Annual: n=105
Interim: n=65
1
2
3
4
5
6
Total test
1st quartile
4th quartiles
Total test
1st quartile
4th quartiles
sample
Early
Late
sample
Early
Late
Constant
-0.008
(-1.380)
(0.168)
0.014
(1.205)
(0.230)
-0.028
(-2.55)
(0.012**)
-0.002
(-0.188)
(0.851)
0.003
(0.102)
(0.919)
-0.002
(-0.112)
(0.910)
SUE
0.038
(6.580)
(0.000***)
0.047
(3.840)
(0.000***)
0.038
(2.967)
(0.004**)
0.057
(4.281)
(0.000***)
0.062
(2.211)
(0.031*)
0.074
(2.580)
(0.012*)
R-Square
0.094
0.125
0.080
0.069
0.072
0.095
F-Statistic
43.29
14.74
9.068
18.35
5.470
6.57
Note: Figure in parenthesis is t-values, significant at 0.05 (*), 0.01 (**), and 0.001 (***) levels
Table 4, (columns 2 to 4) show the regression results of the CAR (50) as dependent variables and SUE
as independent variable for annual announcements. The regressions were done for full test sample,
early announcements sample and late announcement sample. For the annual announcements sample,
the Model 1 of the regression shows that the SUE has a coefficient of 0.38 with a t-statistics of 6.58,
significant at 0.000 level. Next, looking at the regression results of the early annual announcement
sample, the SUE coefficient is 0.47 with a t-statistic value of 3.84, significant at 0.000 level. The
regression results of the late annual announcement sample, the SUE coefficient is 0.038 with a tstatistic value of 2.97, significant at 0.004 level. The R-square values for full test, early and late annual
1235
Proceedings of Applied International Business Conference 2008
announcements sample are 9.4 %, 12.5% and 8.0% respectively. From these results, we can conclude
that early annual announcement sample has received the most vigorous responses from the investors,
whereas the late annual announcement has received less significant responses from investors in term of
the magnitude of the slope of the SUE coefficients and R-square value. Figure 2 plots the respective
curves and shows that the early annual announcements sample has the highest magnitudes and slope as
compare to the full test annual announcements sample and the late annual announcements sample.
CAR
0,05
0,04
0,03
0,02
0,01
0
-0,01
-0,02
-0,03
Total
-0,5
0
0,5
Early
Late
-0,04
-0,05
-0,06
SUE
Figure 2: Plot of CAR and SUE for early, late and total annual announcement sample
0,04
0,03
0,02
0,01
CAR
0
-0,01
-0,5
0
0,5
Total
Early
Late
-0,02
-0,03
-0,04
-0,05
-0,06
SUE
Figure 3: Plot of CAR and SUE for early, late and total interim announcement sample
Table 4, (columns 5 to 7) shows the regression results of the CAR (50) as dependent variable and SUE
as independent variable for interim announcements. The regressions were done for full test sample,
early announcements sample and late announcement sample. For the interim announcements sample,
the Model 4 of the regression shows that the SUE has a coefficient of 0.0.57 with a t-statistics of 4.281,
significant at 0.000 level. Next, looking at the regression results of the early interim announcement
sample, the SUE coefficient is 0.062 with a t-statistic value of 2.211, significant at 0.031 level. The
regression results of the late interim announcement sample, the SUE coefficient is 0.074 with a t1236
Proceedings of Applied International Business Conference 2008
statistic value of 2.58, significant at 0.012 level. The R-square values for full test, early and late annual
announcements sample are 6.9%, 7.2% and 9.5% respectively. These findings suggest that late interim
announcement sample has received the most vigorous responses from the investors, whereas the full
test annual announcement has received less significant responses from investors. Figure 3 plots the
respective curves and the early interim announcements sample has the higher magnitudes as compare to
the full test annual announcements sample and the late interim announcements sample.
Finally, comparing the results for annual and interim earnings announcements, annual earnings
announcements are more significantly priced by investors than the interim earnings announcements.
The firms that announce their annual earnings early received more vigorous responses than the late
annual earnings announced. Similarly, Figure 3 shows that in the case of interim earnings, investors
response to the early interim announcements stronger than the late announcements, however, the slope
or the late interim earnings announcement is steeper than the early interim earnings announcement.
This might be due to the difference in the information content of the annual and interim earnings
announcements. The majority of the annual announcements were made three or four months after the
financial year-ends, whereas the interim earnings announcements were announced mostly within two
and three months of the scheduled date. Furthermore, the annual announcements were mostly audited
announcement, and the interim earnings announcements were all non-audited accounts. Thus, some
users have argued that the interim results announced are too tentative and subjective without a potential
audit opinion hence perceived as less important than annual earnings.
Table 5:
Independent
Variable
Constant
Regression Results For Returns-to-Earnings Relation For Firms announce
earnings early( 1st Quartile) and late (4th Quartile)
Regression Model: CAR i = a1 + a2 SUE i+ a3 De i+ a4 Dl i+ εii
Dependent Variable: Abnormal returns (AR),
Independent Variable: Standardized Unexpected Earnings (SUE)
Annual: n=430
Interim: n=254
Model
1
2
3
4
5
6
-0.015
(-2.563)
(0.011*)
-0.000
(-0.024)
(0.981)
-0.009
(-1.230)
(0.219)
-0.021
(-0.162)
(0.871)
-0.004
(-0.271)
(0.787)
-0.004
(-0.265)
(0.791)
SUE
0.037
(6.394)
(0.000***)
0.037
(6.344)
(0.000***)
0.036
(6.262)
(0.000***)
0.057
(4.214)
(0.000***)
0.057
(4.268)
(0.000***)
0.056
(4.165)
(0.000***)
De
0.034
(2.850)
(0.004**)
0.028
(2.204)
(0.028*)
-0.0002
(-0.008)
(0.994)
Dl
-0.043
(-2.353)
(0.019*)
-0.029
(-1.517)
(0.130)
0.002
(0.074)
(0.940)
0.058
(0.219)
(0.827)
0.007
(0.231)
(0.818)
R-Square
0.107
0.102
0.110
0.068
0.069
0.100
F-Statistic
26.07
24.65
18.20
9.13
9.15
6.08
VIF
1.006
1.010
1.01-1.13
1.007
1.003
Note: Figure in parenthesis is t-values, significant at 0.05 (*), 0.01 (**), and 0.001 (***) levels
1.01-1.20
Table 5 shows the regression results of the CAR (50) as dependent variable and the SUE, De and Dl
(Dummy variables) as independent variables. The variable SUEs are highly significant in all the 6
regression models. In Model 1, the dummy variable De has a t-statistics of 2.85, significant at 0.01
levels. The R-square values have improved from 9.4% (Table 4, Model 1) to 10.7 %. This indicates
that the firms that announce annual earnings early have a positive significant coefficient, which means
there is positive information content beyond unexpected earnings. The Model 2 shows that Dl has a tstatistics of –2.535 significant at 0.05 levels. Therefore, the variable Dl represents firms that announced
1237
Proceedings of Applied International Business Conference 2008
annual earnings late have information content beyond unexpected earnings and the effect is negative to
the earnings-returns relation. Model 3 is the regression of all the independent variables. Variables SUE
and De are highly significant. The R-square values have increased from 9.4% again to 11%. Although
the increase is marginal, however, it does provide evidences that information on the timing of firms that
announce annual earnings have information content beyond earnings
The coefficients for De and Dl are 0.034 and -0.043 respectively. This show that the firms in quartile 1
that announced earnings early have a 3.4% higher returns than the rest, whereas the firms in quartile 4
that announced earnings late has a negative 4.3% in their returns to earnings announced. These results
further confirms the direction and magnitude effect of these annual earnings announcements.
Table 5, (columns 4 to 6) shows the regression results of the CAR (50) as dependent variable and the
SUE, De and Dl (Dummy variables) as independent variables for interim earnings. In these regression
all the coefficients for De and Dl were not significant, therefore, implies that timing of interim earnings
announcements have no value to the investors. There were no significant responses from investors to
the timing of the interim earnings announcements, the investors focused on the magnitude of the
interim earnings per se during announcements. Therefore, there is no information content beyond the
interim earnings announcements for timing the interim earnings releases.
6. Conclusion
This paper investigates the timeliness of the annual and interim earnings reports announcements and
the direction and magnitudes of unexpected earnings and the market reaction to these announcements.
The findings suggests that the direction and magnitudes of the unexpected earnings are different
between the firms that release their annual reports early and the firms that release annual reports late.
Markets reaction to these releases as measured by the risk-adjust cumulative returns also show
significance difference. The effects on the timing of interim earnings announcement however are less
pronounced.
The Malaysian firms announce positive unexpected earnings early and announce negative unexpected
earnings late. Unable to hide bad news due to the mandatory disclosure requirements, Malaysian firms
have turned to delaying their announcement of annual reports with bad news. Alternatively, the
Malaysian firms’ managers are more concerned about internal evaluation that directly effects their
compensation. Therefore, they usually require more time to prepare responses, or announce it together
with good news that comes along. The above literature confirms the timing behavior of Malaysian
firms.
The investors react more vigorously to financial reports that are announced early, and response to the
delay of earnings announcements with less enthusiasm by reacting less vigorously for late earnings
announced. This is consistent with Cheng’s et. al. (2001) evidence that investors price share value
closely to unexpected earnings.
In conclusion, the evidence shows that the direction and magnitudes of earnings affect the time of the
annual reports releases, and the market reacts to the announcements expectedly. It substantiates the
evidence on the timing behavior of Malaysian listed firms in Bursa Malaysia, shows the direction and
magnitudes of earning that effect the timing behavior and investors’ positive reaction to early
announcements. Finally, the results show that investors perceive more information content in annual
earnings announcements compared to interim earnings announcements.
References
Ali, A. (1994). The incremental information content of earnings, working capital from operations, and
cash flows. Journal of Accounting Research, 32, 1, 61-74.
Ali, A., and Pope, P.F. (1995). The incremental information content of earnings, fund flows and cash
flows, The UK evidence. Journal of Business Finance and Accounting, 22, 1, 19-34.
Ariff, M., Loh, A. L. C., and Chew, P. M. K. (1997). The impact of accounting earnings disclosures on
stock prices in Singapore. Asia Pacific Journal of Management, 14, 17-21.
Basu, S. (1997). The conservatism principle and the asymmetric timeliness of earnings. Journal of
Accounting and Economics, 24, 3-37.
Begley, J., and Fisher, P. E. (1998). Is there information in an earnings announcement delay? Review of
Accounting Studies, 3, 347-363.
1238
Proceedings of Applied International Business Conference 2008
Beaver, W. H., Clarke, R., and Wright, W. (1979). The association between unsystematic security
returns and the magnitude of earnings forecast errors. Journal of Accounting Research, 17, 2,
316-340.
Beaver, W. H., Lambert, R., and Morse, D. (1980). The information content of security prices. Journal
of Accounting and Economics, 2, 3-28.
Beaver, W. H., Lambert, R. A., and Ryan, S. G., (1987). The information content of security prices, a
second look. Journal of Accounting and Economics, 9, 2, 139-157.
Bowen, R., Burgstahler, M., and Daley, L. (1987). The incremental information content of accrual
versus cash flows. The Accounting Review, 62 (October), 723-47.
Chambers, A. E., and Penman, S. H. (1984). Timeliness of reporting and stock price reaction to
earnings announcements. Journal of Accounting Research, 22, 21-47.
Chaney P.K. and Jeter D.C. (1993). The effect of size on the magnitude of long-window earnings
response coefficients. Contemporary Accounting Research, 8, 540-560.
Chen, C. R., and Mohan, N. J. (1994). Timing the disclosure of information, management’s view of
earnings announcements. Financial Management, 23, 3, 63-69.
Cheng F.F., M. Ariff and M. Shamsher, (2001). Accounting earnings and share revaluation: Further
exploration. Capital Markets Review, 9, 21-48.
Cheng, C.S.A., Chao, S.L., and Schaefer, T.F. (1996). Earnings permanence and the incremental
information content of cash flows from operations. Journal of Accounting Research, 34, 1,
175-181.
Chen,G. Cheng L,T,W., and Gao N. (2005) Information content and timing of earnings announcements.
Journal of Business Finance and Accounting, 32,1,2, 65-95.
Collins Clubb, D.B. (1995). An empirical study of information content of accounting earnings, funds
flow and cash flow in the U.K. Journal of Business Finance and Accounting, 22,1, 35-52.
Cotter, J. (1996). Accrual and cash flow accounting models: A comparison of the value relevance and
timeliness of their components. Accounting and Finance, 36, 127-150.
Dhaliwal D.S., Lee K.J. and Fargher N.L. (1991). The association between unexpected earnings and
abnormal security returns in the presence of financial leverage. Contemporary Accounting
Research, 8, 20-41.
Fowler, D.J., Rorke, C. H. (1983). Risk management when shares are subjected to infrequent trading.
Journal of Financial Economics, 12, 279-289.
Hagerman, R.L., Zmijewski, M.E., and Shah, P. (1984). The association between the magnitude of
quarterly earnings forecast errors and risk-adjusted stock returns. Journal of Accounting
Research, 22 , 2, 526-540.
Haw, I.M., Qi, D., and Wu, W. (2000). Timeliness of annual report releases and market reaction to
earnings announcement in an emerging capital market: The case of China. Journal of
International Financial Management and Accounting, 11, 2, 109-130.
Jennings, R. (1990). A note on interpreting incremental information content. The Accounting Review,
85, 925-932.
Kross, W. (1982). Profitability, earnings announcement time lags, and stock prices. Journal of Business
Finance and Accounting, 22, 313-328.
Lev, B., (1989). On usefulness of earnings and earnings research - Lessons and directions from two
decades of empirical research. Journal of Accounting Research, 27, 2,153-192.
Opong, K.K. (1995). The information content of interim financial reports: UK Evidence. Journal of
Business and Accounting, 22, 2, 269-279.
Rayburn, J. (1986). The association of operating cash flows and accruals with security returns.
Journal of Accounting Research, 24 (supplement), 112-133.
Schadewitz, H. (1996). Information content of interim earnings components - Evidence from Finland.
Journal of Business Finance and Accounting, 23, 9, 1397-1414.
Sharpe, W.F. (1963). Capital asset prices: A theory of market equilibrium under conditions of risk.
Journal of Finance, XLX, 3, 425-442.
Shamsher M. and Cheng F.F. (2001). The effect of auditor choice on the earnings-to-returns
relationship: The case of Malaysia. Global Business and Finance Review, 6, 2, 53-64.
Sinclair, N. A., and Young, J. C. Y. (1999). The timeliness of half yearly earnings announcements and
stock returns. Accounting and Finance, 6, 32-52.
Swaminathan, S., and Weintrop, J. (1991). The information content of earnings, revenues and
Expenses. Journal of Accounting Research, 29,2, (Autumn), 418-27.
1239