the conservatism in accounting and its effect on earnings response coefficient in tehran stock...
DESCRIPTION
http://www.ijsrpub.com/uploads/papers/IJSRK/IJSRK-2014/Jan/IJSRK-13-229.pdfTRANSCRIPT
International Journal of Scientific Research in Knowledge, 2(1), pp. 28-37, 2014
Available online at http://www.ijsrpub.com/ijsrk
ISSN: 2322-4541; ©2014 IJSRPUB
http://dx.doi.org/10.12983/ijsrk-2014-p0028-0037
82
Full Length Research Paper
The Conservatism in Accounting and Its Effect on Earnings Response Coefficient in
Tehran Stock Exchange Listed Companies
Abbas Ramezanzadeh Zeidi1*
, Zabihollah Taheri2, Ommolbanin Gholami Farahabadi
3
1Department of Accounting, Neka Branch, Islamic Azad University, Neka, Iran
2Department of Accounting, Payamenour University, Sari, Iran
3Department of Accounting, Payamenour University, Neka, Iran
*Corresponding Author: E-mail: [email protected]
Received 11 November 2013; Accepted 27 December 2013
Abstract. The main purpose of this research is to investigate the relationship between conservatism and earnings response
coefficient. For conservatism the book-to-market ratio and for measurement of earnings management Ohlson's valuation model
is used. The statistic population of the research includesthe listed companies in Tehran Stock Exchange and the samples
include 154 companies over the period of 2007 to 2012. Using the panel regression analysis, several models have been
estimated and the hypotheses have been examined. Results show that there is negative and significant relationship between
conservatism and earnings response coefficient.
Keywords: Conservatism, Earnings Response Coefficient, Earnings per Share
1. INTRODUCTION
Conservatism has been a controversial concept since
accounting started as a science, and it has a significant
role in the practice of accounting. Basu believes that
conservatism has affected both theoretical and applied
accountings for centuries. Historical records of the
early 15th century concerning partnership and
collaborative trade trades show that accounting has
been conservative in Medieval Europe. Conservatism
is one of the dominant features of financial reports
which have attracted more attention due to scandals
(in companies like Enron and WorldCom). Some
recent studies, such as Watts (2003), Roychowdhury
and Watts (2005), Lafond and Watts (2006), are
specifically focused on conservatism. The notion of
conservatism was introduced when the balance sheet
was the most important, if not the only, financial
report, and the components of earnings and other
operational results were less often presented out of the
companies.
In securities market, investment decisions are
affected by factors like news media, analysts, and
financial reports of companies. Therefore, different
responses of investors towards earnings information
lead to different responses by the market. But what are
the reasons behind the different responses of the
market? Scott, in his book titled financial accounting
theory, presents several reasons including systematic
risk (beta), quality of earnings, continuation of
profitability or loss, opportunities for investment
growth, and capital structure. Thus, in earnings
response coefficient two variables must be
determined; abnormal returns and unexpected
earnings. New securities market values that react to
new information can be observed. In fact, the response
of the securities market for a number of unexpected
earnings is higher for some companies. Earnings
response coefficient is different responses of the
market to the earnings information. In other words,
earnings response coefficient measures the
unexpected output in response to the unexpected
earnings reported by the companies which publish the
securities (Khoshtinat et al., 2008).
One of the factors that may lead to a different
response of the market to earnings information is
conservatism. Many researches have been done on the
responses of the market to earnings information.
Based on such researches, since the notions of
conservatism and earnings response coefficient are
applied in accounting differently for financial
reporting, and each one can affect the quality of
financial reporting, and hence affect the efficiency of
investment market and the behavior of investors,
creditors, and in general behavior of all who use the
financial lists, it seems that investigating the relation
Zeidi et al.
The Conservatism in Accounting and Its Effect on Earnings Response Coefficient in Tehran Stock Exchange Listed
Companies
82
between these two factors can contribute to the
literature of this issue.
In this study, the researcher tries to answer the
question of whether conservatism in accounting
approaches affects the earnings response coefficient,
and what is the response of earnings coefficient of
each share in such companies.
2. BACKGROUND AND LITRATURE REVIEW
Many researches have been done concerning the
earnings response coefficient and the factors affecting
it. According to these studies, the investors’ decisions
lead to the market response. The new values of the
securities market which respond to the new
information can be observed. The first evidences of
the securities market response to publishing and
announcement of earnings were provided by Ball and
Brown (1986). In addition to demonstrating the
market response to publishing the earnings according
to the prediction, they measured the amount of the
response in forms of mean and average. They showed
that investors’ response to the companies with good
news resulted in unexpectedly positive outcome and
to the companies with bad news led to unexpectedly
negative outcome. Ball and Brown applied the
average of unexpected securities in their research.
This average can cover a high amount of deviation
from the mean. Thus, the unexpected outcome is
probably at the average level for some companies and
below the average for some others.
Vikil (1990) investigated the effect of audit change
on the earnings quality and the earnings response
coefficient. He hypothesized that earnings response
coefficient in the companies that had changed their
audit would face a significant change; however, he
could not prove his hypothesis based on the statistic
results.
In a research trying to provide a way to interpret
the market from the published and announced
information of earnings, Ceris and Samer (2005),
directly related earnings response coefficient with
continuation of profitability trend. This research
differs from other researches concerning the effect of
continuation of profitability trend on the earnings
response coefficient in that it considered the
continuation of unexpected earnings trend as the
effective variable rather than the continuation of
profitability trend. Moreover, in interpretation of the
earnings response coefficient, rather than dividing it
into two levels, stable and instable, it is divided into
several new levels. The findings proved the relation
between the variables of the research.
Tosh, Zayang and Jian (2005), tried to show that
the companies which disclose the increases due to
keeping in profits and incomes have a more earnings
quality and higher earnings response coefficient in
comparison with those companies that disclose only
increases due to keeping in profit. Their research
findings proved their assumptions.
In a research, Louis (2011) investigated the
relation between conservatism accounting and the
value of kept cash money with that of abnormal
outcome of a company. He found that conservatism
accounting affects positively the relation between kept
cash money and future operation of the company. In
addition, the results prove that conservatism
accounting is related to optimal use of available cash
of the company. The results also supported the view
that conservatism accounting can be are placement for
foreign supervision, resulting in expense decrease.
Abed et al. (2012) investigated a level of
conservatism in accounting and its relation with
earning management. The results of their research
showed that levels of conservatism differ in different
companies. They also found that conservatism and
earnings managements are negatively correlated.
Khoshtinat and Joshghani (2006) examined the
effects of financial leverage on earnings response
coefficient in Tehran Stock Exchange listed
companies. The objective of their research was to
answer the question of whether investors and analysts
consider the capital structure of the companies and
how much leverage they were in their response to
good and bad news resulting from publishing
accounting information, and whether their response
affected such news. In this research, they employed
the balance sheet approach to measure financial
leverage. In the balance sheet approach two
definitions are given for leverage: the ratio of debt to
assets and the ratio of debts to the share holders’ right,
both of which have been considered in this research.
Investigation of the only hypothesis by regression
analysis during the period of 2000-2004 shows that in
the first definition of leverage in the entire sample and
in the second definition of leverage in the high level
of leverage, there are negative relations between the
financial leverage and the earnings response
coefficient. However, in the first definition of
Leverage in the low level and in the second definition
in the entire sample in the low level no significant
relation was observed between the financial leverage
and the earnings response coefficient.
Ahmadi (2008) studied the relation between the
earnings quality and earnings response coefficient. In
this study the qualitative characteristics of financial
information are used to evaluate the earnings quality.
In this research, firstly, the dimensions of relevance
and reliability benefits have been separately estimated
using time series analysis and combined data methods.
Companies were classified on the basis of high and
low earnings quality based on cluster analysis. In
International Journal of Scientific Research in Knowledge, 2(1), pp. 28-37, 2014
03
order to test the hypothesis Ohlson’s model
(regression price-profit) was used. The findings show
that the earnings response coefficient as well as the
explanatory power of the regression (cost-benefit) is
significantly higher in portfolio of companies with
high quality of earnings than companies with low
quality of earnings. Thus, both hypothesis of the
research were proved.
2.1. The Statistic Population
In the present study, to test the hypotheses classified
financial data of listed companies in Tehran Stock
Exchange has been used. Samples were limited to the
companies with the following specifications:
1) The end of the financial year of the company was
the end of Iranian calendar year.
2) The company did not have a change of the financial
year in two succeeding years.
3) The financial data of the company for the study
period was accessible.
4) The company was not one of the investment
companies, leasing and banks due to the nature of
their activities.
Considering the mentioned limitations, 154 companies
were selected for the study.
2.2. The Research Hypothesis
Based on the preliminary studies, the research
hypothesis is as follows:
There is a significant correlation between
conservatism and earnings response coefficient.
3. METHODOLOGY
This research is based on descriptive-exploratory
approach. It studies the present situation, describing it
systematically, and examines characteristics and
features of the present situation in relations between
the variables. This type of research is not only
practical in the area of decision-making and planning
but is also fundamental since it leads to scientific facts
based on inductive method.
3.1. The Operational Definition of Variables and
the Way They are calculated
Dependent Variable (Earnings Response Coefficient),
The earnings response coefficient measures the
unexpected yield of the market in response to the
unexpected component of reported earnings by a
company that has securities sites. In other words, the
earnings response coefficient measures the sensitivity
of the market to declaration of profit by the slope
coefficient of the regression of abnormal returns and
unexpected earnings (Scott, 2003). In this research,
the dependent variable is the earnings response
coefficient. The earnings response coefficient, which
measures the reaction of the market to changes of
profit of each share and is derived from a regression
of price and earnings per share, accounts for the way
the information regarding profit can affect the stock
price. The assessment model, based on Ohlson’s , is as
follows. Following Barth et al. (1999), the researcher
added two variables, namely risk and growth, to the
equation so that the effects of other variables can be
controlled. The desired variable in this research is the
EPS coefficient.
Price = δ0 +δ1 BVE + δ2 EPS + δ3 (EPS * DE) + δ4
(EPS *EVAR) + ψ
Price: The price of each share at the end of the
financial year
BVE: Book Value of each share
EPS: Earnings per share before unexpected items
DE: Debt to equity ratio (risk)
EVAR: The change percentage in earnings of per
share
3.2. The Independent Variable
Conservatism is one of the dominant features of
financial reporting which have attracted more
attention in recent years due to financial scandals (in
companies like Enron and WorldCom). The
independent variable in this research is conservatism.
According to previous studies (Ahmad et al, 2002;
Zhang, 2007; Lobo et al, 2008), the ratio of the book
value to the market value of equity is considered as
the representative of conservatism. The ratio of book
value to market value of equity being less than one is
an indication of accounting conservatism.
3.3. The Control Variables
3.3.1. Size
Zimmerman (1983) states that big companies are
more conservative due to more political sensitivity
(Political Cost Theory). In the present research, the
natural logarithm of total assets of the end of period is
considered as the indication of the company size.
3.3.2. Operation
Operation, measured through ROE, is calculated by
dividing the earnings, after tax deduction, on the total
shareholders’ equity at the end of the period.
3.3.3. Financial Leverage
Accounting methods are related to financial leverage
Zeidi et al.
The Conservatism in Accounting and Its Effect on Earnings Response Coefficient in Tehran Stock Exchange Listed
Companies
03
since one of the criteria which is considered by the
creditors in Iran, i.e. banks, is the ration of debts of
the company. Therefore, the higher the ratio of debt of
a company is, the less intention it will show to use
conservatism methods. Consequently, company
managers are expected to apply less conservative
methods in their financial statements in order to
reduce the risk of rejection of loan applications and to
avoid incurring higher interest costs. This variable is
defined as total debts to total assets at the end of
period. This ratio is calculated by dividing total debts
to total assets.
3.4. Earnings Management
According to previous studies, discretionary accruals
in a company represent earnings management.
Residual values of the modified Jones’s model of
accruals, that is, discretionary accruals are employed
in this research as the criterion of earnings
management.
3.5. Analyses Concerning the Main Model of the
Research
The hypothesis of this research concerns the relation
between conservatism and earning response
coefficient. This relation is also considered in the
main model of the research. To examine such a
relation, in this research, Ohlson’s general model of
assessment was employed. To examine the main
model of the research, the interactive variable of
conservatism index (the ratio of book value to market
value) and earnings per share is included in the main
model. In this research, to control the effects of some
factors according to previous studies, the researcher
will consider the variables of the ratio of debt to
shareholders’ equity and percentage of change in
earnings per share, size, yield of shareholders’ equity,
earnings management as control variables, the impact
of which will be assessed as interactive relationships.
3.6. Descriptive Statistics of the Main Model
Variables
The descriptive statistics of the variables include
central, dispersion, and distribution indexes. In this
research, information related to mean and median
from central index, standard deviation related to
dispersion index, and stretching and skewing related
to distribution index are presented. In addition, in this
figure, Jarque -Bera statistic and the related level of
significance are presented for examining the normality
of the variables distribution. The descriptive
statistics of the main model variables are given in
table (1):
Table 1: Descriptive Statistics of the Main Model Variables
EM ROE SIZE EVAR DE CONS EPS BV P
-4.50E-12 0.405161 13.38333 11.26447 2.063388 0.880815 639.5525 1885.271 4394.933 Mean
0.003025 0.261357 13.17000 -7.996947 1.730000 0.550846 468.5200 1304.230 2458.000 Median
2.004014 83.21339 19.62000 12526.78 303.3700 131.8100 13448.11 211071.0 58414.00 Maximum
-4.414097 -55.91304 9.820000 -34126.57 -201.1100 -30.02000 -3891.480 -21855.54 150.0000 Minimum
0.233599 5.445384 1.483872 1524.559 15.02313 4.854001 1170.477 7914.886 5685.128 Standard
Deviation
-6.818365 4.407300 0.808653 -11.82005 3.696714 22.24455 3.038666 22.18468 4.270578 skewing
154.5776 131.4621 4.018428 296.2156 245.1163 591.4461 27.01835 556.7610 30.60352 stretching
857948.6 624521.4 137.1351 3291906. 2236983. 13218915 23555.15 11688980 31761.20 Jarkko-
Bra
0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 probabilit
y
Observations show that the average price of each
share of the sample companies is about 4395 Rials.
The book value per share is about 1885 Rials in
average, and the earnings per share is average about
600 Rials. These three are the basic variables of
Ohlson’s assessment model which all have positive
skewing and stretching. Positive stretching indicates
that the curve of variables distribution is longer than
normal distribution. As it is also observed from the
Jarque-Bera and the related level of significance, the
three variables do not have a normal distribution.
Observations indicate that the conservatism index
of the sample companies is about 0.88 in average. As
it was already mentioned, in Jain and Rezaee’s (2004)
belief, the ratio of book value to market value of
equity being less than one is an indication of
accounting conservatism. Since this proportion is less
than one in the sample companies, in these companies
there is conservatism accounting. In half of the
companies the ratio of book value to the market value
is more than 0.55, and in the other half this ratio is
less than 0.55. The debt of the sample companies is
averagely about two times of their shareholders’
equity. The mean of change percentage of earnings
per share shows that the earnings per share of the
companies have changed nearly 11.26 units during the
International Journal of Scientific Research in Knowledge, 2(1), pp. 28-37, 2014
08
period of this research. The yield of shareholders’
equity of the research samples is about 40 percent.
The discretionary accruals, which indicate earnings
management, is -4.50E-12. The minus mark shows a
negative relation between earnings management and
earnings response coefficient.
4. RESULTS AND DISCUSSIONS
4.1. Examining Correlation between the Main
Model Variables
In this section, using Pearson's correlation coefficient,
existence and direction of correlation between the
variables of the model of response coefficient is
examined. The correlation coefficient matrix is given
in table (2).
Table 2: Correlation coefficients between the variables of the main model
Price Book Value
Earnings
per Share
Conservatis
m
Ratio of
Debt Growth Size
Yield of
Shareholder
s’ Equity
Earnings
Management
Price 1.000000
-----
Book Value 0.022718 1.000000
0.5157 -----
Earnings per
Share
0.683421 0.005685 1.000000
0.0000 0.8708 -----
Conservatism -0.066695 0.293781 -0.046933 1.000000
0.0561 0.0000 0.1791 -----
Ratio of Debt -0.007256 -0.003611 0.015985 -0.004290 1.000000
0.8355 0.9177 0.6474 0.9023 -----
Growth -0.168266 0.013576 -0.113545 0.025005 0.029123 1.000000
0.0000 0.6977 0.0011 0.4743 0.4046 -----
Size 0.001632 -0.001908 -0.019292 0.003455 -0.020919 -0.008114 1.000000
0.9627 0.9565 0.5810 0.9213 0.5495 0.8164 -----
Yield of
Shareholders’
Equity
0.053068 0.002673 0.032689 -0.008871 -0.021218 -0.002386 -0.040396 1.000000
0.1287 0.9390 0.3496 0.7997 0.5438 0.9456 0.2476 -----
Earnings
Management
0.052717 -0.021509 0.317746 -0.028300 0.019435 -0.029490 -0.007237 0.009058 1.000000
0.1312 0.5383 0.0000 0.4180 0.5782 0.3987 0.8360 0.7955 -----
As it is seen in this table, considering the level of
significance related to correlation coefficient between
the variables, the intensity of the relationship between
the explanatory variables of the main model is weak
or average, and a strong correlation between the
explanatory variables is not observed. Therefore,
simultaneous entry of variables in the research model
would not interfere in terms of linear time.
4.2. Tests Related to Main Models of the Research
Test of the main model is performed at 95% certainty
level based on multiple variables regression, and in
estimating the general model combined data analysis
will be used. In this analysis, first to choose a suitable
model, Chow test (for choosing a model with or
without effects) and Hausman test (for choosing a
model with fixed effects) are used. For examining
significance of the regression model Fisher’s statistic,
for examining the significance of the regression model
coefficients Student’st statistic, for examining
autocorrelation between observations Durbin-Watson
statistic, and for examining explanatory power of the
model the adjusted coefficient of determination
statistics are used.
4.3. Tests Related to the Model of Correlation
between Conservatism Index and Earnings
Response Coefficient
To choose a suitable model, Chow test and Hausman
test were used. The results of these tests are given in
table (3). Considering the F statistic of Chow test and
the amount of probability (less than 0.05) the model
with effects is suitable. According to the 2
of
Hausman test and the amount of probability (less than
0.05), the model with fixed effects is suitable.
Zeidi et al.
The Conservatism in Accounting and Its Effect on Earnings Response Coefficient in Tehran Stock Exchange Listed
Companies
00
Table 3: The Results of Choosing a Model for Model Test of Relation between Earnings Management and Earnings Response
Coefficient
Test Type Test Statistic Amount of
Statistic
Freedom
Degree
Probability
Chow Test F 5.161335 (153.748) 0.0000
Hausman
Test
2 226.473606 3 0.0000
Table 4: Estimating the model of relation between conservatism index and earnings response coefficient
Dependent Variable: P
Explanatory Variables Coefficients Standard Error Statistict Probability
C 3591.950 59.15273 60.72331 0.0000
Book Value 0.088617 0.031100 2.849427 0.0045
Earnings per Share 1.298114 0.087276 14.87372 0.0000
Interactive relationship between
conservatism and earnings per share
-0.109625 0.042242 -2.595162 0.0096
adjusted coefficient of determination
statistic
0.809490 Fisher’s statistic 25.62276
Durbin-Watson’s Statistic 1.872307 Probability of Fisher’s Statistic 0.000000
P = 3591.950 + 0.088617*BV + 1.298114*EPS - 0.109625*CONS*EPS + [CX=F]
The results of estimating the model by fixed effects
method are given in table (4). Analyzing the results of
model estimating, the followings can be concluded:
the t statistic and probability (less than 0.05) indicate
the significance of both book value per share and
earnings per share. Moreover, the width of source of
the model affects the model positively and
significantly. Considering the fact that the variable
coefficient of earnings per share in the model is
1.298114, the market price of each share responds
positively and significantly to the changes of earnings
per share. In other words, the earnings response
coefficient is positive.
In this model, correlation between conservatism
and earnings response coefficients is performed
through examining the interactive variable. The t
statistic and the probability (less than 0.05) indicate
the significance of the interactive variable. Of course,
the negative coefficient of the interactive variable
indicates a negative and significant correlation
between the index of conservatism and earnings
response coefficients.
In the first model of earnings response coefficient,
the variable coefficient of earnings per share in the
model is 2.002890. In other words, earnings response
coefficient is positive, and the market price of each
share responds positively and significantly to changes
of earnings per share. In the model of correlation
between conservatism index and earnings response
coefficient, the variable coefficient of earnings per
share in the model is 1.298114. In other words,
earnings response coefficient is positive and the
market price of each share responds positively and
significantly to changes of earnings per share.
Decrease in coefficient of earnings per share
represents the impact of conservatism interactive
variable on earnings response coefficient, a fact which
is in accordance with the research findings. The
adjusted R2 statistic of the model indicates that nearly
80 percent of changes in the market price of each
share can be explained by the explanatory variables,
while the adjusted R2 statistic of the first model,
without the presence of the interactive variable of
earnings management is about 37 percent. This
increase of approximately 43 percent in adjusted
coefficient of determination is another reason for the
interactive role of conservatism in the model, and it
can account for high reaction of the market price of
each share to the changes of earnings per share.
Durbin-Watson statistic of the model is 1.872307, and
this rejects the assumption of autocorrelation between
the elements of the model.
4.4. Tests Related to the Model of Correlation
between Conservatism Index and Earnings
Response Coefficient at the Presence of Control
Variables (Ration of Debt and Growth)
To choose a suitable model, Chow test and Hausman
test were used. The results of these tests are given in
table (5). Considering the F statistic of Chow test and
the amount of probability (less than 0.05) the model
International Journal of Scientific Research in Knowledge, 2(1), pp. 28-37, 2014
03
with effects is suitable. According to the X2
of
Hausman test and the amount of probability (less than
0.05), the model with fixed effects is suitable.
The results of estimating the model by fixed effects
method are given in table (6).
Table 5: The Results of Choosing a Model for Model Test of Correlation between Earnings Management and Earnings
Response Coefficient at the Presence of the Control Variables
Test Type Test Statistic Amount of
Statistic
Freedom
Degree
Probability
Chow Test F 7.317205 (153.731) 0.0000
Hausman
Test
2 246.543710 5 0.0000
Table 6: Estimating the model of correlation between conservatism index and earnings response coefficient at the presence of
the control variables of leverage and growth
Dependent Variable: P
Explanatory Variables Coefficients Standard Error Statistict Probability
C 3671.761 52.32307 70.17480 0.0000
Book Value 0.022108 0.009979 2.215518 0.0270
Earnings per Share 1.081997 0.078471 13.78846 0.0000
Interactive correlation between
conservatism and earnings per share
-0.094261 0.034362 -2.743143 0.0062
Interactive correlation between ratio of
debt and earnings per share
0.012191 0.003992 3.053633 0.0023
Interactive correlation between growth
and earnings per share
-0.000128 1.14E-05 -11.20461 0.0000
adjusted coefficient of determination
statistic
0.834353 Fisher’s statistic 28.81274
Durbin-Watson’s Statistic 1.958772 Probability of Fisher’s Statistic 0.000000
P = 3671.761 + 0.022108*BV + 1.081997*EPS - 0.094261*CONS*EPS
+ 0.012191*DE*EPS - 0.000128*EVAR*EPS + [CX=F]
Analyzing the results of model estimating, the
followings can be concluded: the t statistic and the
probability (less than 0.05) indicate the significance of
both book value per share and earnings per share.
Considering the fact that the variable coefficient of
earnings per share in the model is 1.081997, the
market price of each share responds positively and
significantly to the changes of earnings per share. In
other words, the earnings response coefficient is
positive.
The variable coefficient of earnings per share in
this model decreased in comparison with the previous
models. This means that earnings response coefficient
in the model with the interactive role of conservatism
and with the control variables is decreased compared
with the previous situation in which the control
variable were not present, indicating that the control
variables affects the model. The adjusted R2 statistic
of the model indicates that nearly 83 percent of
changes in the market price of each share can be
explained by the explanatory variables. The adjusted
coefficient of the model with interactive role of
conservatism is higher than the previous models.
Durbin-Watson statistic of the model is 1.958772, and
this rejects the assumption of autocorrelation between
the elements of the model.
4.5. Tests Related to the Model of Correlation
between Conservatism Index and Earnings
Response Coefficient at the Presence of Control
Variables (Ration of Debt, Percentage of Earnings
Changes, Size, Profitability, and Earnings
Management)
Analyzing the results of model estimating, the
followings can be concluded: the t statistic and the
probability (less than 0.05) indicate the significance of
both book value per share and earnings per share.
Considering the fact that the variable coefficient of
earnings per share in the model is 1.196885, the
market price of each share responds positively and
significantly to the changes of earnings per share. In
other words, the earnings response coefficient is
positive.
The variable coefficient of earnings per share in
this model has increased in comparison with the
previous models. This means that earnings response
coefficient in the model with the interactive role of
conservatism and with the control variables (yield and
earnings management) is again increased compared
with the previous situation, showing the role of the
newly added control variables. Among the newly
added control variables, the size is not significant, the
control variable of yield has a positive and significant
Zeidi et al.
The Conservatism in Accounting and Its Effect on Earnings Response Coefficient in Tehran Stock Exchange Listed
Companies
03
effect on the model, and the earnings management has
a negative and significant effect. The adjusted R2
statistic of the model indicates that nearly 89 percent
of changes in the market price of each share can be
explained by the explanatory variables. The adjusted
coefficient of the model with interactive role of
conservatism is higher than the previous models.
Durbin-Watson statistic of the model is 1.921116, and
this rejects the assumption of autocorrelation between
the elements of the model.
Table 7: Estimating the model of correlation between conservatism index and earnings response coefficient at the presence of
control variables
Dependent Variable: P
Explanatory Variables Coefficients Standard Error Statistict Probability
C 3498.928 39.85992 87.78062 0.0000
Book Value 0.086725 0.036095 2.402697 0.0166
Earnings per Share 1.196885 0.297220 4.026930 0.0001
Interactive correlation between
conservatism and earnings per share
-0.124159 0.049813 -2.492503 0.0129
Interactive correlation between ratio of
debt and earnings per share
0.021150 0.005418 3.903452 0.0001
Interactive correlation between growth
and earnings per share
-0.000131 8.93E-06 -14.66161 0.0000
Interactive correlation between size and
earnings per share
0.010274 0.020782 0.494354 0.6212
Interactive correlation between yield
and earnings per share
0.042912 0.019164 2.239226 0.0255
Interactive correlation between earnings
management and earnings per share
-0.620493 0.083784 -7.405852 0.0000
adjusted coefficient of determination
statistic
0.892406 Fisher’s statistic 43.44999
Durbin-Watson’s Statistic 1.921116 Probability of Fisher’s Statistic 0.000000
P = 3498.928 + 0.086725*BV + 1.196885*EPS - 0.124159*CONS*EPS
+ 0.021150*DE*EPS - 0.000131*EVAR*EPS + 0.010274*SIZE*EPS
+ 0.042912*ROE*EPS - 0.620493*EM*EPS + [CX=F]
5. CONCLUSION
In this research, first the interactive variable of
conservatism was fitted without control variables. The
results show that the earnings response coefficient is
positive and its amount is less than the primary
situation without the interactive variable of earnings
management, indicating that the interactive variable of
conservatism affects the model. In the second stage,
the interactive variables of risk and growth in form of
interaction were inserted into the model together with
the interactive variable of conservatism. In this case,
the findings showed a positive earnings response
coefficient, yet the amount was less than the first case
in which no control variable was present. This
indicates that the control variables of risk and growth
influence the amount of earnings response coefficient.
In the last stage, the interactive variables of size, yield
of shareholders’ equity, and earnings management
were inserted into the model. The findings in this
case, too, show a positive earnings response
coefficient. This time, also, the earnings response
coefficient was higher in comparison with the
previous situation in which only two variables of
growth and risk were present, indicating the effect of
new control variables on the changes of earnings
response coefficient. In general, it can be concluded
that there is a negative and significant correlation
between conservatism and earnings response
coefficient, and adding the interactive variable of
conservatism will lead to reduction of earnings
response coefficient in the model.
Suggestions for Future Studies:
1- It is suggested that the role of earnings smoothing
on statement of earnings response coefficients are
examined.
2- It is suggested that time-series properties of
earnings, including earnings persistence, earnings
predictability and the like should be considered in the
evaluation of earnings response coefficients.
3- It is suggested that earnings response coefficient is
separately investigated in different bourse industries.
REFERENCES
Abed S, Al-Badainah J, Serdaneh JA (2012). The
Level of Conservatism in Accounting Policies
and Its Effect on Earnings Management.
International Journal of Economics And
Finance, 4(6): 78-85.
International Journal of Scientific Research in Knowledge, 2(1), pp. 28-37, 2014
03
Ahmed AS, Billings BK, Morton RM, Harris MS
(2002). The Role of Accounting Conservatism
in mitigating bondholder-shareholderconflicts
over dividend policy and in reducing debt costs.
The Accounting Review, 77 (4): 867–890.
Bae B, Sami H (2005). The effect of potential
Environmental Liabilitieson Earnings Response
Coefficients. Journal of Accounting, Auditing
&Finance, 43-70.
Barth ME, Beaver WH, Hand JRM, Landsman WR
(1999) Accruals, Cash Flows, and Equity
Values. Review of Accounting Studies, 4: 205-
229.
Basu S (1997). The Conservatism Principle and the
Asymmetric Timeliness of Earnings. Journal of
Accounting and Economics, 3-37.
Ballsam S, Krishnan J, Yang J, (2003). Auditor
Industry Specialization and the Earnings
Response Coefficient. Auditing: A journal of
practice & Theory.
Belkaoui AR (2002). The Effects of Multinationality
on EarningsResponse Coefficients. Managerial
Finance, 28: 97-106.
Chang A, Shiva K (2010). Earnings Management and
Earnings Predictability. www.ssrn.com,
working paper.
Collins D, Kothari S (1989). An Analysis of
International and Cross-Sectional Determinants
of Earnings Response Coefficients. Accounting
and Economics, 11: 143-18.
Dechow PM, Skinner DJ (2000). Earnings
management: Reconciling the views of
accounting academics, practitioners, and
regulators. Accounting Horizons, 14(June):
235-250
Hong T, Wong S, Perceived T (1993). Quality
Auditor and Earnings Response Coefficient",
Accounting Review, 68(2): 346-366.
Jain P, Rezaee Z (2004). The Sarbanes- Oxley Act of
2002 & Accounting Conservatism.
www.ssrn.com, Working Paper series.
Khani A (2007). The relationship between accounting
profit and accounting abnormal returns profit in
Tehran Stock Exchange. Journal of Faculty of
Administrative Sciences and Economics,
University of Tehran, 19(1).
Khoshtinat M, Josheghani FH (2008). Accounting
Studies, 17: 1-26.
Lafond R (2005). The influence of ownership
structure on earnings conservatism and the
informativeness of stock prices: An
international comparison. On line:17.
http://www.ssrn.com;
Lobo GJ, Parthasarathy K, Sivaramakrishnan S
(2008). Growth, Managerial Reporting
Behavior, and Accounting Conservatism.
Working paper Series. University of Houston.
Roychowdhury S, Watts R (2007). Asymmetric
Timeliness of Earnings, Market to Book and
conservatism in financial reporting. Journal of
Accounting & Economics 44: 2-31.
Perols JL, Lougee A (2011). The relation between
earnings management and financial statement
fraud. Contents lists available at ScienceDirect,
39 -53.
Rahmani A, Bashirimanesh N, Shahrokhi S (2012).
Accounting Knowledge , 10: 29-50.
Saghafi A, Bolo G (2009). Cost of shareholders salary
and earnings attributes. Accounting Research,
2: 44-59.
Watts. Ross L (2003). Conservatism in Accounting,
Part II : Evidence and Research Opportunities,
www.ssrn.com,Working Paper.
Zhang F (2007). Accruals, Investment, Growth, and
The accrual Anomaly. The Accounting Review.
82(5): 1333-1363.
Zeidi et al.
The Conservatism in Accounting and Its Effect on Earnings Response Coefficient in Tehran Stock Exchange Listed
Companies
03
Abbas Ramezanzadeh Zeidi received his MA in Accounting from Tehran Branch, Islamic Azad
University. Currently, he is PhD Candidate at AMU India. He has more than 20 papers in the referees
journals and conferences. He is faculty member of Neka Branch, Islamic Azad University.
Zabihollah Taheri received his MA in Accounting from Tehran Branch, Islamic Azad University. He is
faculty member of Payamenour University, Sari, Iran.
Ommolbanin Gholami Farahabadi received her MA in Accounting from Payamenour University,
Behshahr, Iran. She is faculty member of Payamenour University, Neka, Iran.