EPS as a measure of intercompany performance: Philippine evidence.
Cudia, Cynthia P. ; Manaligod, Gina T.
INTRODUCTION
International Accounting Standards (IAS) 1 states that the
objective of financial reporting is to provide information that is
useful to a wide range of users. Stice, Stice and Skousen (2007) explain
that an investor uses this financial information in making credit and
investment decisions. Investors gauge how well a company performs in
comparison with other companies. Profitability is one of the basic
indicators of the soundness of a business entity. However, it is not
enough to know whether net income is increasing or decreasing. Investors
are concerned with how income is changing relative to certain factors
such as firm size.
The analysis of the financial performance of a business becomes
more meaningful when profit is scaled against acceptable measures of
firm size such as total assets and sales. Ratios such as return on
assets (ROA) and return on sales (ROS) are commonly used to evaluate the
results of business operations. Another popular measure of income
performance is Earnings per Share (EPS). It is a ratio which
incorporates both net income and outstanding ordinary shares in the
measurement of profit performance. In this tool, the number of ordinary
shares is proposed as a measure of firm size. This financial ratio is
very important to most users, particularly the ordinary shareholders
because EPS information is helpful to them in evaluating the return of
their investment and risk of a corporation (Nikolai and Bazley, 2003).
Questions now arise. Can EPS be used in making comparisons across
companies? Can outstanding ordinary shares be used as a measure of firm
size? To help illustrate this concern is the case of two local companies
in the Philippines, namely, Pilipinas Shell Petroleum Corporation and TI
Inc. These are two large corporations with similar asset size. Both
companies have total assets of approximately P59 billion as of December
31, 2007. For the year ending 2007, Pilipinas Shell and TI Inc. posted
net income of P 6.4 million and P5 million respectively. Given this
information, the net income of Pilipinas Shell exceeded the net income
of TI Inc. by P1 million. Scaling these profit figures by total assets
resulted in Pilipinas Shell and TI Inc. providing returns on asset
investment of 10.61% and 8.55%, respectively. These percentages
reinforce the early observation that Pilipinas Shell is more profitable
than TI Inc. But in terms of EPS, Pilipinas Shell registered EPS of
P9.19 while TC Inc. posted an EPS of P524.64. This contradicts the
previous analysis that Pilipinas Shell performed better than TI Inc. It
can be seen therefore that when EPS is used as a measure of financial
performance, the landscape of the profitability picture changes. This
change can be attributed to the number of outstanding ordinary shares
which is the measure of firm size used in the EPS ratio. In this
particular case, Pilipinas Shell has 68 times more outstanding ordinary
shares than TI Inc.
The analysis of the income performance of these two corporations
reveals the possible drawbacks of comparing profitability among firms
using EPS. The landscape of a firm's profit picture changes when
the number of ordinary shares is taken into consideration. It is in this
light that this research paper is undertaken. This study determined if
the number of ordinary shares is an acceptable measure of firm size that
can be used in evaluating the reasonableness of using EPS as a tool of
measuring financial performance across corporations in the Philippines.
BACKGROUND LITERATURE
International Accounting Standards (IAS) 33 defines EPS as the
amount of net income attributable to every outstanding ordinary share
during a period of time. The standard states that the purpose of
presenting EPS is to provide financial statement users with information
on the performance of a single entity. Creditors and investors find this
measurement tool as an effective way of evaluating the income
performance of each outstanding ordinary share. This is the only
financial indicator that incorporates both net income and share
investment in the computation making EPS a significant measure of
financial performance.
Jordan, Stanley and Robert (2007) of the University of Southern
Missouri studied if the number of ordinary shares outstanding represents
a reasonable measure of company size and a legitimate way to scale
earnings. Data was collected on 300 publicly traded companies. From each
size group based on market capitalization, 100 companies were randomly
selected. The study showed that EPS comparisons among large publicly
traded companies may be appropriate. However, it should not be made
among small publicly traded firms because the number of ordinary shares
outstanding represents a poor scaling measure for entity size. The
researchers concluded that accounting professors should refrain from
teaching EPS as a tool for making inter company performance comparisons.
Lie and Lie (2002) evaluated multiples that are used to estimate
company value. They found that asset multiple generally generates more
precise and less biased estimates than sales and earnings multiples. The
research showed that the accuracy and bias of value estimates as well as
the relative performance of the multiples vary greatly by company size,
profitability and intangible value in the company.
De Berg and Murdoch (1994) conducted an empirical investigation of
the usefulness of earnings per shares disclosures. The study examined
whether both primary earnings per share and diluted earnings per share
have the potential to provide financial statement users with information
that is useful. The research concluded that primary and diluted earnings
per share contain essentially the same information. Because of this, the
study further concluded that it was quite improbable these data could be
utilized as separate independent variables in a predictive decision
model.
An examination was made by Graham and King (2000) regarding the
relation between stock prices and accounting earnings and book values in
Indonesia, Korea, Malaysia, Philippines, Taiwan and Thailand. They found
differences across the six countries in the explanatory power of book
value per share and EPS for firm values. Explanatory power for Taiwan
and Malaysia was relatively low while that for Korea and the Philippines
was relatively high.
Hodgson and Stevenson-Clark (2000) used Australian data to
determine whether stock returns, earnings and cash flows are important
in addressing the issue of whether accounting data provide value
relevant information. They observed that a non linear functional
relation provides greater explanatory power for both earnings and cash
flow. This result was consistent for smaller firms but contrary for
larger firms.
Huff and Harper (1999) concluded that there are systematic
differences among liquidity and solvency measures for small companies
versus large companies. The means of both the current ratio and the debt
ratio were large for the small companies. There was evidence indicating
the wide variability of values for both current ratios and debt ratios
among small companies compared to large companies. In all cases, the
variances were much larger for small companies, which suggests more
comparability among large companies than small companies.
Ten popular local and financial accounting textbooks are reviewed
as part of the related literature of this study. Valix and Peralta
(2009) gives a thorough discussion of IAS 33 but it fails to mention if
this ratio can be used in making inter firm comparisons. Robles and
Empleo (2006) indicate that the objective of EPS information is to
improve historical comparisons among different enterprises in the same
period and among different accounting periods for the same enterprise.
Padilla and Flores (2006) simply is silent regarding the reasonableness
of using EPS in making comparisons across companies. Chalmers, Mitrione,
Fyfe, Weygandt, Kieso and Kimmel (2007) indicate in their book that the
only meaningful EPS comparison is an intra-entity trend comparison
stating that inter company comparisons are not meaningful because of
wide variations in the number of issued shares among companies.
Spiceland, Sepe, Tomassini (2001) writes that summarizing performance in
a way that permits comparisons is difficult. Deegan (2007) stresses that
care must be taken when comparing various entities' basic and
diluted EPS because calculations accounting profits are heavily
dependent upon professional judgment. Stice, Stice and Skousen (2007),
Dyckman and Davis (2001), Nikolai and Bazley (2003) and Kieso, Weygandt
and Warfield (2008) fail to mention if EPS can be used in making inter
company comparisons. Only one out of the ten books reviewed made a clear
statement that EPS comparison is meaningful only when it is
intra-entity.
HYPOTHESES
This study validates Jordan, Stanley and Robert (2007) using
Philippine data. This paper attempts to find out whether EPS can be
taught as a means of comparing inter company performance in the
Philippines. The studies of Lie and Lie (2002), Graham and King (2000)
and Hodgson and Stevenson-Clark (2000) establish that there are
relationships among asset values, stock prices, earnings and cash flows.
Hence, the variables used in this study include alternative measures of
company size such as total assets, sales, net income and shares
outstanding.
Deberg and Murdoch (1994) showed that primary or basic EPS is
essentially the same as diluted EPS. As such, this study uses basic EPS.
Finally, Huff and Harper (1999) proved that there are differences in
liquidity and solvency ratios based on company size. Hence, this study
divides the sample-companies between small and big companies to test the
following hypotheses:
1. There is a significant difference between ROS of big firms and
small firms in the Philippines.
2. There is a significant difference between EPS of big firms and
small firms in the Philippines.
3. The number of ordinary shares is a legitimate measure of firm
size in the Philippines.
4. There is a strong linear relationship between outstanding
ordinary shares and two other measures of firm size, namely, total
assets and sales, of firms in the Philippines.
METHODOLOGY
This study utilized data collected from the 2007 financial
statements of 233 companies belonging to the top 300 corporations in the
Philippines by gross revenues. The companies belong to the large
category as defined by the Philippine government. Under the Small and
Medium Enterprise Development (SMED) Council Resolution No. 01 Series of
2003 issued by the Department of Trade and Industry, a large enterprise
is defined as a business whose total assets must have a value of above
P100 million. The largest asset size in the sample-companies is P 1,083,
005 million while the smallest asset size is P489 million. The
sample-companies are divided into two groups using the median of total
assets. All companies with asset values of P6,981 million and above are
referred to as Group 1 or Big Firms while all those corporations with
asset values below P6,981 million are referred to as Group 2 or Small
Firms. There are 117 companies in Group 1 and 116 firms in Group 2.
T-test was performed on the means of EPS and Return on Sales (ROS)
of the two sample groups to determine whether these are statistically
different from one another. Return on sales is computed by getting the
quotient of net income over sales. ROS is used as a basis of comparing
EPS because ROS is a common measure of income performance. Furthermore,
statistical tools such as correlation analysis, regression and ANOVA
were employed to determine if the number of ordinary shares is a
legitimate measure of firm size, and to examine the strength of relation
between outstanding ordinary shares to total assets and sales.
RESULTS AND DISCUSSION
Using the methodology described in the previous section of this
paper, the following discussion of data analysis is presented:
Table 1 provides summary measures for ROS and EPS for the big firms
and the small firms. In terms of ROS, the group of big firms registered
a mean ROS of 24.14% while the small firms posted a mean ROS of 3.68%
only. This indicates that the big firms performed better than the small
firms. The ROS mean difference between the two sample groups of.20461 is
statistically significant, with a p-value of 2.98E-09 at an a level of
.05.
However, the mean analysis of EPS is not consistent with the
results of the ROS mean analysis. EPS analysis showed that the mean EPS
for big and small firms are P196.20 and P198.55 respectively. These
figures show that small firms fared slightly better than the big firms.
Table 1 shows that the EPS mean difference between the two groups at a
level of .05 produced a p-value of .9798, which indicates that the EPS
mean difference between the two sample groups is not statistically
significant.
If EPS is an acceptable alternate measure of earnings scaled based
on company size, then the comparison of EPS between the big and small
firms should have followed a pattern similar to the comparison of ROS
means for the two sample groups. However, results of this study showed
otherwise. The analysis of mean ROS indicates that there is a large
difference in profitability between the big and small firms. Whereas,
the analysis of mean EPS suggests that there is only a slight difference
in profitability between the two sample groups. This finding already
proves that the use of EPS in making inter company comparisons may not
yield the same results if ROS is used.
To determine whether the number of outstanding ordinary shares is a
suitable measure of firm size, the strength of relation between the
number of shares and other accepted measures of firm size such as total
assets and sales is tested using correlation analysis.
For the big firms, the more traditional measurement of firm size
like total assets and sales exhibited weak relationship with a
correlation coefficient of .378. Similarly, the relationship between
total assets and shares has a computed correlation coefficient of .008
showing very weak relationship between them. On the other hand, sales
and shares are negatively correlated with a correlation coefficient of
-.014.
For the small firms, measures of firm size like total assets, sales
and shares showed weak relationships with correlation coefficients of
.321 and .373. Moreover, there is weak correlation between sales and
shares with a coefficient of .103 only.
From an overall perspective with all the firms taken into
consideration, the total of assets is weakly correlated with sales
having a correlation coefficient of .433. With a coefficient correlation
of .099 and .076 respectively, there are also weak relationships between
total assets and shares and between sales and shares.
If the number of shares is a reliable proxy of firm size, then it
must have a strong relationship with the other common measures of size.
Correlation analysis shows that the number of shares has a weak
relationship with total assets and sales regardless of category.
Therefore, using outstanding ordinary shares as an acceptable measure of
firm size is not established in this research.
In addition to the correlation matrix, regression and ANOVA is done
to determine the strength of the statistical relationships between the
number of shares and the other measures of entity size such as total
assets and sales.
Tables 5 and 6 presents ANOVA regression models with the number of
shares regressed on each of the other measures of firm size to wit,
total assets and sales.
For big firms, overall test of significance yielded p-value of
.9786, which implies that there is no linear relationship between shares
and the two variables. The results of ANOVA test are confirmed by the
regression output. Taken individually, the variable total assets with a
coefficient of 397.70 and p-value of .8828 indicate that there is no
linear relationship between number of shares and total assets. Likewise,
the variable sales, with coefficient of -1891.79 and .8486 p-value,
indicates no linear relationship between sales and number of shares.
For small firms, overall test of significance yielded p-value of
.0002, which implies that there is a linear relationship between shares
and the two variables. Likewise, the ANOVA test was performed to confirm
the regression output. The variable total assets with a coefficient of
91,538 and a p-value of .0001 suggest that there is a linear
relationship between the number of shares and total assets. However, the
variable sales with a coefficient of -2,133 and a .8375 p-value mean
that there is no linear relationship between number of shares and sales.
The regression analysis computed R2 of0.000 and 0.140 for big and
small firms, respectively. For both sample groups, the model showed no
predictive power.
CONCLUSIONS
Based on the analysis, the mean ROS of the big firms is
statistically different from the mean ROS of the small firms. On the
other hand, the mean EPS of the big firms is not statistically different
from the mean EPS of the small firms. Since no pattern is established
between ROS and EPS, then it is concluded that EPS is not an acceptable
tool in making inter company comparisons of profitability.
The correlation matrices indicate that, in general, the number of
outstanding ordinary shares has no strong relationship with the other
common measure of firm size such as total assets and sales. This
conclusion is true for big firms, small firms and when all the firms are
taken as one single group. It is therefore concluded that the number of
shares is a weak predictor of firm size.
Statistical tests using ANOVA and regression analysis revealed that
the number of shares of the big firms does not show a linear
relationship with their total assets and sales. This might imply that
any change in total assets or sales would not result to a corresponding
level of change in number of shares. Assets and sales are not predictors
of the change in number of shares for the group of big firms.
On the other hand, ANOVA table for the small firms showed that
there is a linear relationship between the number of shares and the two
variables, total assets and sales. Moreover, results of the regression
analysis concerning the small firms differ from those of the big firms.
For the group of small firms, regression results revealed that there is
a linear relationship between number of shares and total assets. The
very small p-value of .0001 of total assets was offset by the high
0.8375 p-value of sales. This leads to the conclusion that there is no
linear relationship between number of shares and sales. This further
indicates that a change in total assets signals either an increase or
decrease in the number of shares. However, change in sales for small
firms does not predict a level of change in the number of shares.
IMPLICATIONS AND RECOMMENDATIONS
The objective of International Accounting Standards (IAS) 1 is to
provide guidance on the preparation of general purpose financial
statements to allow comparability of reports both across time and across
companies. IAS 33 is silent regarding the use of EPS in making
comparisons among business entities. Many financial accounting textbook
authors, both foreign and local, fail to mention if EPS can be used in
making inter company comparisons. There is no warning to readers
regarding the possible pitfalls of comparing EPS among firms. The
results of this study showed that the number of shares does not proxy
for firm size for both the big and small firms. It is therefore not a
viable measure for comparing performance among corporations. For
academic purposes especially in the Philippines, accounting professors
should teach EPS with an emphatic statement that such tool cannot be
used for inter company comparisons. It is recommended that EPS not be
presented to students as an appropriate tool of evaluating inter-company
profit performance. Or at the very least, students should be made aware
of the pitfalls of using EPS in making comparisons across companies.
REFERENCES
Chalmers, Keryn, Mitrione, Lorena, Fyfe, Michelle, Weygandt, Jerry,
Kieso, Donald & Kimmel, Paul. (2007/ Principles of Financial
Accounting. John Wiley & Sons Ltd. (Australia)
De Berg, Curtis, Brock, Murdoch. (Spring 1994). An Empirical
Investigation of the Usefulness of Earnings Per Share Disclosures.
Journal of Accounting, Auditing and Finance (Boston, USA).
Deegan, Craig. (2007). Australian Financial Accounting. McGraw-Hill
Australia Pty Ltd.
Dyckman, Thomas, Davis, Charles, Dukes, Roland. (2001).
Intermediate Accounting Fifth Edition. McGraw-Hill Publishing.
Graham, Roger, King, Raymond (2000). Accounting Practices and the
Market Valuation of Accounting Numbers: Evidence from Indonesia, Korea,
Malaysia, Philippines, Taiwan and Thailand. The International Journal of
Accounting.
Hodgson, Allan, Stevenson-Clark, Peta. (2000). Earnings, Cash Flows
and Returns: Functional Relations and the Impact of Firm Size.
Accounting and Finance (Australia).
Huff, Patricia Lee, Harper, Robert Jr., Eikner, Elaine. (June
1999). Are There Differences in Liquidity and Solvency Measures Based on
Company Size?. American Business Review (USA).
International Accounting Standard (IAS) 33
Jordan, Charles, Clark, Stanley, Smith, W. Robert. (2007). Should
Earnings Per Share (EPS) Be Taught as a Means of Comparing Inter Company
Performance? The Heldref Publications.
Kieso, Donald, Weygandt, Jerry, Warfield, Terry. (2008).
Intermediate Accounting Twelfth Edition. John Wiley & Sons (Asia)
Pte. Ltd.
Lie, Erik, Lie, Heidi. (2002). Multiples Used to Estimate Corporate
Value. Financial Analysts Journal (USA).
Nikolai, Loren, Bazley, John. (2003). Intermediate Accounting Ninth
Edition. Thomson Learning.
Padilla, Nicanor, Flores, Erminio. (2006). Financial Accounting
Theory of Accounts. GIC Enterprises and Co. Inc.
Robles, Nenita, Empleo, Patricia. (2006). Intermediate Accounting
Volume 3. Mutual Books, Inc.
Spiceland, David, Sepe, James, Tomassini, Lawrence. (2001).
Intermediate Accounting Second Edition. McGraw-Hill Publishing.
Stice, James, Stice, Earl, Skousen, Fred. (2207). Intermediate
Accounting Sixteenth Edition. Thomson Learning.
Valix, Conrado, Peralta, Jose. (2009). Financial Accounting Volumes
1 and 2. GIC Enterprises & Co., Inc. (Manila, Philippines).
Cynthia P. Cudia, De La Salle University
Gina T. Manaligod, De La Salle University
Table 1: Return on Sales (ROS) and Earnings Per Share (EPS) for Big
and Small Firms
Return on Sales Earnings Per Share
(ROS--%) (EPS--P)
Big Small Big Small
Mean 24.14% 3.68% P196.20 P198.55
Sample Variance 11.86 0.90 618,589.70 382,931.50
Standard Deviation 34.43 9.47 786.50 618.81
N 117 116 117 116
df 231 231
Mean difference .20461 -2.34922
t-statistic 6.17 -0.03
p-value (two-tailed) 2.98E-09 .9798
Table 2: Correlation Matrix--Big Firms
Variable Total Assets Sales Shares
Total Assets 1.000
Sales 0.378 1.000
Shares 0.008 -0.014 1
Sample size 117
Critical value .05 (two-tail) [+ or -].182
Critical value .01 (two-tail) [+ or -].237
Table 3: Correlation Matrix--Small Firms
Variable Total Assets Sales Shares
Total Assets 1.000
Sales 0.321 1.000
Shares 0.373 0.103 1.000
Sample size 116
Critical value .05 (two-tail) [+ or -] .182
Critical value .01 (two-tail) [+ or -] .238
Table 4: Correlation Matrix--All Firms
Variable Total Assets Sales Shares
Total Assets 1
Sales 0.433 1
Shares 0.099 0.076 1
Sample size 233
Critical value .05 (two-tail) [+ or -] .129
Critical value .01 (two-tail) [+ or -] .168
Table 5: Regression and ANOVA--Big Firms
Regression Analysis--All Possible Regressions
117 observations
G1 Shares is the dependent variable
G1 TA G1 s Adj [R.sup.2] Cp p-
Sales [R.sup.2] value
0.8835 3654759674.48 0 0 1.022 0.8835
0.935 3654996528.73 0 0 1.037 0.935
0.8828 .8486 3670403115.880 .000 .000 3.000 0.9786
Regression Analysis
[R.sup.2] 0
Adjusted [R.sup.2] 0 n 117
R 0.019 k 2
Std. Error 3670403115.88 Dep.Var. G1 Shares
ANOVA Table
Source SS df MS
Regression 582254009026347100 2 291127004513174050
Residual 1535791929769240500000 114 13471859033063502000
Total 1536374183778270500000 116
Source F Pvalue
Regression 0.02 0.9786
Residual
Total
Regression Output
Variables Coefficients Std. error t (df=114) p-value
Intercept 1807248525.23 409402080.676 4.414 0
G1 TA 397.6991 2692.4781 0.148 0.8828
G1 Sales -1891.7865 9888.9572 -0.191 0.8486
Table 6: Regression and ANOVA--Small Firms
Regression Analysis--All Possible Regressions
116 observations
G2 Shares is the dependent variable
G2 TA G2 Sales s Adj [R.sup.2] Cp p-
[R.sup.2] value
0 370897330.13 0.132 0.139 1.042 0
0.0001 0.8375 372465206.73 0.124 0.14 3 0.0002
.2714 397640072.23 .002 .011 17.931 0.2714
Regression Analysis
[R.sup.2] 0.14
Adjusted [R.sup.2] 0.124 n 116
R 0.374 k 2
Std. Error 372465206.73 Dep.Var. G2 Shares
ANOVA Table
Source SS df MS
Regression 2542025461915590500 2 1271012730957790200
Residual 15676527315330405000 113 138730330224163000
Total 182185527772460050000 115
Source F Pvalue
Regression 9.16 0
Residual
Total
Regression output
Variables Coefficients Std. error t (df=114) p-value
Intercept -107437393.536 90727191.4125 -1.184 0.2388
G2 TA 91538.0367 22246.2436 4.115 0.0001
G2 Sales Variables 10376.1477 -0.206 0.8375
