920
ijt
t
ijt
j
ijt
ijt
j
ijt
j
ijt
s
yeardummie
ROA
Assets
netsales
Assets
TCA
H
E
J
E
D
'
1
1
1
1
(3)
Where, Sales are defined as gross sales and other operating revenue less discounts returns and
allowances. Lagged ROA, the operating income after taxes relative to total assets is included to control for
firm performance. This model estimates the pooling data of accounting data required by the SET during
1998-2006. The parameters achieved from equation (3) will be used to estimate the EPTCA in the
following equation (4):
dummies
year
ROA
Assets
AR
sales
net
Assetsi
EPTCA
t
ijt
ijt
j
jt
ijt
.
ˆ
ˆ )
( ˆ
1 ˆ
1
1
E
J
E
D
'
'
(4)
Where
AR
'
denotes the change in account receivables and all other variables as defined earlier.
Event study of earnings announcement
The reaction of the market on earning announcement could be observed through the abnormal
returns which represent the result of new information at time of the announcement. The firms are divided
into two groups - politically connected firms and ordinary firms. As for the measurement of abnormal
returns, the standard event-study methods
1
are applied to examine the stock price responses to earning
announcements. Day 0(t=0) in event time is defined as the date of announcement of the firms’ annual
financial statement and the event window is 80 days around the event date, 59 days before the event date
and 20 days after.(t=-59 to t=5) The Capital Asset Price Model(CAPM) which indicates the prices of the
securities in an efficient market will be used to estimated the expected return of each stock(
it
R
) through
the following equation;
it
mt
i
i
it
u R
R
E
D
The excess returns are then calculated using the raw excess return method displayed in the following
equation:
mt
it
it
R R AR
After averaging the abnormal returns across all firms, the statistical significance is tested using the t-
statistic. The similar steps are applied to both the ordinary firms and political connected firms to see
whether the investors react in the same way toward the firms’ earnings announcement. However, as
investors have been focused on the profits of the firms or the earning itself, the reaction toward good news
(the positive earnings) and the bad news (negative earnings) must be filtered out of the market reaction.
1
See Dodd and Warner (1983) for detailed description of event-study methods and statistic tests
.
