Earnings response coefficients in the Greek market.
Maditinos, Dimitrios I. ; Sevic, Zeljko ; Stankeviciene, Jelena 等
1. Introduction
A fundamental issue of economics, finance and accounting involves
the relation between the firm's reported earnings and its stock
returns (Kormendi, Lipe 1987). Standard valuation models assume that
price is the discount present value of future expected dividends or
future cash flows. It is commonly assumed that, over long periods, re
ported accounting earnings are directly related to future dividends and
cash flows. Ball, Brown (1968) numerous studies have attempted at
identifying whether reported earnings contain information used by the
market for assessing the value of the common stock of the firm.
In the late 1980s, researchers started investigating a new
area--the earnings response coefficient (ERC) that is theoretically
defined as "a change in the price induced by a one-dollar change in
current earnings" (Collins, Kothari 1989) and typically measured as
a slope coefficient in a regression of stock returns on unexpected
earnings (Markowitz 1952, 1959). While the studies on the average
price-to-earnings ratio (Stankeviciene, Gembickaja 2012) are
concentrated on market reactions to earnings announcements, the studies
on the earnings response coefficient are more interested in the nature
of information about reported earnings and how they are related to firm
valuation (Kormendi, Lipe 1987).
The main objective of this study is to analyse the relationship
between accounting data and market price returns of the companies listed
in the Athens Stock Exchange (ASE). More specifically, the article
describes if there is a statistically significant earnings response
coefficient of the companies listed in the ASE conducting annual
cross-sectional and intertemporal regression analysis. The paper
endorses and advances the methodology used by Kothari, Sloan (1992) and
Jindrichovska (2001). The major findings of this study may contribute to
various groups of people such as investors, corporations, regulators,
educators and researchers.
The present study is organised as follows: section two consists of
literature review discussing various relevant issues of research on the
earnings response coefficient. The purpose of this work is to provide
the basic theoretical and a detailed review of the earnings response
coefficient (ERC). Moreover, it presents an empirical foundation for
other studies and their implications. Section three examines the issues
of research design and provides the research method, including a
detailed discussion of the model employed. Section four contains the
description of data and core results obtained from statistical analysis
and focuses on answering the research methodology developed in the
previous chapter. Section five discusses and summarises the findings of
the study, including limitation on the results and suggestions for
future research.
2. Framing issues: theoretical considerations
There are many approaches for how accounting data affect market
price returns. Ball, Brown (1968) documented a positive statistical
association between earnings surprises and stock returns around earnings
announcements. A voluminous body of research (Beaver et al. 1980; Brown
et al. 1987; Beaver 1989) has examined the role of accounting earnings
in financial markets. In rather influential papers, that prompted
further research, Easton, Zmijewski (1989), Collins, Kothari (1989) and
Kormendi, Lipe (1987) empirically tested the last implication. Ohlson,
Schroff (1992) confirmed that if investors used other information than
that about earnings and dividends alone, there was a reason to prefer
one specification over the other. Easton, Harris (1991) used a different
method and examined earnings as an explanatory variable for return and
confirmed the relation ship between the level of earnings (scaled by
price) and stock returns at the beginning of the period. The main
difference in this study is that it has incorporated the level of and
changes in earnings rather than only a change in earnings. Kothari
(1992) and Kothari, Sloan (1992) also examined the strength of the
relationship between price and earnings. On the contrary to the previous
studies, they deflated earnings by the beginning-of-the year share
price, including three leading period returns. They explored
price-earnings regressions when 'prices lead earnings'.
Collins et al. (1994) incorporated up to three years of future
earnings in their returns-earnings regressions and found the levels of
explained return association that were higher when compared to
regressions that only used contemporaneous earnings.
When employing the methodology proposed by Kothari, Sloan (1992)
and a sample of UK companies, Donnelly, Walker (1995) supported the view
that an increase in the ERCs realised by incorporating leading prices in
return earnings relation were not specious. Kothari and Zimmerman (1995)
provided an explanation of why return models were commonly preferred to
price models. Nevertheless, return models have less serious econometric
problems than price models (Liu, Thomas 2000). Consistent to Kothari,
Zimmerman (1995) the results obtained by Martikainen et al. (1997) and
Dumontier, Labelle (1998) in Finland and France respectively were
presented.
Hayn (1995) noted that losses were very important when estimating
return-earnings relation, because they were not expected to continue
forever, since shareholders had a liquidation option. When loss
observations are excluded, the association between returns and earnings
becomes much stronger. This is supported by Finnish data collected by
Martikainen et al. (1997) and Kallunki, Martikainen (1997).
Amir, Lev (1996) reported that the value of financial information
was largely irrelevant in returns-earnings regressions, whereas that of
nonfinancial information was highly relevant. Jermakowicz,
Gornik-Tomaszewski (1998) evidenced a significant association between
stock returns and accounting earnings and concluded that the annual
earnings were an important element of the valuation process of a firm.
Jindrichovska (2001), following the methodology proposed by Kothari
and Sloan (1992), reported that one-leading-year returns were as
important as contemporaneous returns in terms of their sensitivity to
annual changes in earnings. Jarmalaite (2002), on the basis of the
methodology developed by the same Kothari, Sloan (1992), analysed the
relationship between accounting data and stock price returns in the
stock markets of Lithuania, Latvia and Estonia. The results suggested
that the relationship between returns and earnings in Latvia seemed to
be very similar to Estonia and Lithuania showing the weakest and Estonia
showing the highest value relevance.
Myring et al. (2003) indicated that both Australian and United
States markets reacted quite quickly to earnings releases. Similar
results were reported by Liu and Thomas (2000) who investigated the role
of analysts' earnings forecasts for explaining returns in Sweden.
Suwardi (2009) discovered that in Jakarta Stock Exchange (JSX) the book
value of net assets appeared to have a stronger relationship with the
market value compared to the US studies that used similar estimated
models.
Kousenidis (2005) employed the Easton, Harris (1991) and found that
the explanatory power of earnings for contemporaneous stock returns was
not significant. Papadaki, Siougle (2007) confirmed a negative
price-earnings ratio regarding firms suffering from loss and a positive
price-earnings ratio as regards profitable firms.
Maditinos et al. (2007) provided evidence that there was a relation
between EPS and stock market returns. The results concerning ROI and ROE
were not significant. Moreover, they showed that the pair-wise
regression of EPS and ROI best explained stock market returns in Greece,
compared to the results provided by the combinations of EPS and ROE as
well as by ROI and ROE.
Dimitropoulos, Asteriou (2009) concluded that the price model
produced less biased ERCs than the return model but faced econometrical
problems. The results showed the increased ability of the price and
return models to explain better earnings-return relationship by
providing highly significant earnings response coefficients.
Furthermore, after correcting value-irrelevant noise in earnings, the
return model yields highly significant ERCs. These results are
consistent to Kothari, Zimmerman (1995), Martikainen et al. (1997) and
Dumontier, Labelle (1998).
3. Methodology
According to Kothari, Sloan (1992) and Jindrichovska (2001),
investors anticipate the numbers of future earnings at least some
periods ahead. This is what is meant by "prices lead
earnings". The reason for such a phenomenon is that the accounting
system does not reveal information about future earnings in a timely
manner while investors instantly adjust prices to their expectations of
future profitability. This idea means that investors not only use
information on past or current earnings numbers for stock valuation, but
are assumed to use more information.
Kothari, Sloan (1992) presented a method for leading period returns
to control the biased coefficients of return models and extended the
return measurement window to several years before the fiscal year of
interest, which resulted in higher earnings response coefficients.
Figure 1 illustrates this idea.
This paper is aimed at examining the degree of the relationship
between market prices and accounting data at both contemporaneous
one-period return-earnings relation and leading period returns in the
regression model. We examine if returns measured over one to four
leading periods contain information about changes in annual earnings.
Using the methodology suggested by Kothari, Sloan (1992), the degree of
the relationship between price relatives (one plus the buy-and-hold
return) and earnings-to-price ratio (earnings yield) is tested using a
quarterly, yearly and intertemporal sample. It is employed the earnings
level like Kothari, Sloan (1992) and Jindrichovska (2001) did, rather
than change, deflated by price as the explanatory variable in the price
earnings regression, which is motivated by the random walk time series
property of annual earnings Ohlson (1995), and the evidence in Easton,
Harris (1991) and Maditinos et al. (2007).
First, the model using contemporaneous earnings and prices is
employed:
[P.sub.it]/[P.sub.it-[tau]] = [a.sub.i] + [[gamma].sub.1i]
[X.sub.it]/[P.sub.it-[tau]] + [[epsilon].sub.it], (1)
where [X.sub.it] is accounting earnings over period t,
[P.sub.it]/[P.sub.it-[tau]] is one plus the buy-and-hold return
inclusive of dividends over the period from the end of t-[tau] to the
end of t. On the right-hand side, fraction [X.sub.it]/[P.sub.it-[tau]]
represents the earnings-to-price ratio (earnings yield). The numerator
consists of accounting earnings per share before extraordinary items for
firm i over the period from the end of t-[tau] to the end of t. Earnings
are divided by the stock price at the beginning of period t-[tau] which
is the same as the price at the end of the previous period.
[[gamma].sub.1i] is the ERC. The approach in this study has been applied
to yearly as well as quarterly estimating windows.
[FIGURE 1 OMITTED]
Notes: Measurement intervals of earnings and returns in lead-lag
price-earnings regressions based on two-year buy-and-hold returns,
inclusive of dividends, regressed on annual earnings deflated by price
at the beginning of the return measurement interval. The return
measurement interval consists of the contemporaneous and one leading
year.
Return observations are overlapping
After contemporaneous estimates, the estimation interval is
lengthened and new coefficients are evaluated over a longer measurement
interval up to four periods. In the next step, the independent variable
is modified and earnings are summed for (t - [tau]) periods.
The model then changes to
[P.sub.it]/[P.sub.it-[tau]] = [a.sub.i] + [[gamma].sub.1i]
[X.sub.it-[tau],t/[P.sub.it-[tau]]] + [[epsilon].sub.it].
As [tau] increases, it is more likely that information content
reflected in [X.sub.it] will be contained in the return over period
t-[tau] to t. As a result, [[tau].sub.1i] (ERC) is expected to approach
its predicted value of (1 + r). As Kothari and Sloan (1992) states, the
multi-year return over period t-[tau] to t reflects more information
than that reflected in the one year earnings, [X.sub.it]. Furthermore,
since this uncorrelated information resides in the dependent variable,
it does not bias the estimated ERC, though the explanatory power of
([R.sup.2]) will be less than one.
4. Empirical results
The sample is selected from the total number of Greek firms listed
in the ASE. The sample period spans from 1998 to 2008 (we have chosen
the period between 1998 and 2008 because it covers the market boom, the
great recession in the period 1999-2000 and the integration of Greek
economy to the European monetary system in 2001, since the Greek stock
market has demonstrated considerable financial stability). The initial
sample is eliminated by excluding all companies falling into the
financial sector. 10 companies have been excluded from annual analysis
because of a lack of annual observations. There are 245 publicly traded
companies in the sample having a different number of participating years
for each of them. These companies have given a total of 2.166 to 1.441
firm year-observations. The accounting data have been obtained from the
ASE data bank and the ICAP Company. Security prices have been extracted
from the ASE database.
Descriptive statistics on earnings yield and price relatives
computed from yearly data for the period 1998-2008 at measurement
intervals ranging from one to four years are reported in Table 1. The
number of observations varies from 2.166 to 1.441. With reference to the
collected data, sample mean, standard deviation, median, minimum and
maximum values are obtained and summarized.
The first and second panel contains descriptive statistics on
earnings yield. In the first panel, both mean values and medians
increase with time. Similar results were found by Kothari, Sloan (1992)
and Jindrichovska (2001). However, the earnings yield of four year
measurement window decreases. Moreover, variability measured by standard
deviation is much higher than that presented in the study by Kothari and
Sloan (1992). In Panel 2, a non-cumulative earnings yield shows an
increasing mean and median over two years and a decreasing mean and
median over four years. Jarmalaite (2002) reported similar results over
a two year period. Our results are not consistent to those presented in
the studies by Kothari, Sloan (1992) and Jindrichovska (2001). For
example, Kothari, Sloan (1992) state that the mean and median earnings
yield increases monotonically along with [tau].
Descriptive statistics on price relatives, including dividends on
various measurement interval samples, is reported in the third panel of
Table 1. An important point is that high variability that increases
together with a measurement interval of two years, encounters a decrease
in the following years. Table 2 displays the estimated ERC of the first
cross-sectional regression model of the intertemporal (all years)
sample. The estimated ERC shows higher sensitivity with leading period
returns. The earnings response coefficient measured using ordinary least
squares with common intercepts increases from 0.139 (t = 12.914) for
contemporaneous estimates to 0.266 (t = 19.003) when returns over one
leading-year is included to 0.317 (t = 22.867) for two leading-years and
0.321 (t = 22.136) for the four-year estimation window. The results
indicate improvement in the ERC over the estimated period using annual
data. All coefficients are significantly different from zero at the
level of 5%. The explanatory power measured by adjusted [R.sup.2]
increases along with the window and ranges from 0.071 to 0.254.
Similar conclusions are found in Collins, Kothari (1989), Easton et
al. (1992) and Kothari, Sloan (1992). However, coefficients are rather
small compared with Kothari, Sloan (1992) and Jindrichovska (2001) and
their variance is not high. The basic conclusion is that the estimated
ERCs are significant and show a stronger relation with stock returns as
an increase in the leading periods. These results suggest that the model
well describes the relationship between price relatives and accounting
earnings in the ASE.
Table 3 shows the results received from the annual cross-sectional
sample in the following way: in the case of contemporaneous estimates,
eight out of eleven regressions (except the years 2000, 2005 and 2008)
are statistically significant at the level of 0.05 according to
t-statistics. The ERCs range from 0.119 (t = 5.591) in 2001 to 0.048 (t
= 2.207) in 2007. The Adj. [R.sup.2]s range from 0.137 to 0.016 each
year respectively. For one leading year, all regressions are significant
at the level of 0.05. The ERCs vary between 0.261 (t = 5.659) in 2001 to
0.068 (t = 7.271) in 2008. In the four year estimation window, five out
of eight coefficients (except years 2000, 2005 and 2008) are significant
at the level of 0.05 according to t-statistics, ERCs range from 0.147 (t
= 3.64) to 0.281 (t = 7.556) and Adj. [R.sup.2]s increase in all cases.
Negative coefficients for a one-year window are puzzling and
statistically insignificant and Adj. [R.sup.2]s are quite low. The
results in this section suggest that the estimates of the earnings
response coefficient, using Kothari, Sloan (1992) methodology, increase
with the inclusion of leading-period returns in accordance with the
tested hypothesis. Furthermore, they behave as the hypothesis
predicts--increase along with the estimation window.
After the completion of annual estimates, quarterly results were
assessed. The estimation of ordinary least squares is used for
estimating the common slope coefficient. However, the quarterly model
did not yield high explanatory power.
Table 4 reports the results (all years) of the cumulative model
(3.2) that represents the cross-sectional OLS regression of aggregated
earnings divided by the price of the return measurement window and price
relatives over the period of one up to four years. According to Kothari
and Sloan (1992) and Jindrichovska (2001), when earnings are aggregated
and the earnings measurement window increases, the variation of results
should be smaller. The results show that all ERC coefficients are
significant at the level of 0.05 and grow as the estimation window
increases up to three periods. In the case of contemporaneous estimates,
explanatory power is very low (the same as in Table 4) and the ERC is
0.139 (t = 12.914). When the period of leading years is examined, it
rises to 0.275 (t = 18.160) and 0.328 (t = 21.704) for two and three
years respectively. The results are consistent to those identified by
Kothari and Sloan (1992) and Jindrichovska (2001), despite higher ERCs.
Generally, the model indicates the increased ability of the return model
to explain the relationship between published earnings and stock prices
with the inclusion of one leading-period return, as the earnings
response coefficient substantially increases and converges toward the
predicted value. However, the relation does not hold in the longer run
([tau] = 4) as the ERC drops to 0.321 (t = 21.111).
Table 5 shows the results of the annual cross-sectional sample. In
all cases, except years 2001 and 2008, the ERC is highly significant at
the level of 0.05 and follows a similar pattern as that for the full
year sample. As the hypothesis predicts, the ERC increases as returns
from leading years are included. Furthermore, the Adj. [R.sup.2]
improves when the periods are led forward in all cases. These results
are consistent with those found in Kothari and Sloan (1992) and
Jindrichovska (2001) and indicate that the cumulative model explains the
relationship between price relatives and earnings as the theoretical
framework suggests.
As the obtained results indicate, the first model using annual data
and the cumulative model well explain the relationship between earnings
and returns. The explanatory power of each model increases along with
the measurement window, and therefore provides useful information to ASE
investors.
5. Concluding remarks
The main aim of the paper was to analyze the relationship between
accounting earnings and returns on the Greek market. It has been
conducted using data on published earnings and stock prices covering the
period from 1998 to 2008 and focusing on earnings response coefficients
estimated using returns measured over a long interval (one quarter up to
four years). The estimates have been calculated for the annual
cross-sectional and intertemporal sample relying on the methodology
proposed by Kothari and Sloan (1992). According to empirical results,
the estimates of the earnings response coefficient are sensitive to the
inclusion of leading-period returns.
As for the first model, empirical evidence from cross-sectional
regression analysis suggest that information provided by
earnings-to-price ratios is of some value relevance for explaining
market price returns in ASE, i.e. the estimated earnings response
coefficient is found to be significant to the leading periods and
increased when more leading periods are included. The explanatory power
of the model increases from 7.1 to 25.4%. Consistent with the studies by
Kothari and Sloan (1992) and Jindrichovska (2001), the estimated ERC is
significant and shows a stronger relation with stock returns as the
leading periods increase, i.e. prices contain information about the
earnings of listed firms. The obtained results are similar to those
received annually and recommend that prices anticipate earnings and
therefore provide useful information to ASE. For a short estimation of
the window of up to three-quarters, the model did not behave as
expected.
The relationship between contemporaneous earnings-to-price ratios
and price relatives is not statistically significant. The quarterly
model (all years and annual) has revealed that, with the inclusion of
more than one leading period, the ERC increases as the theory predicts.
However, the explanatory power of the model remains rather low in order
to be considered adequate. As the results depict, the cumulative model
provides higher and more significant ERCs when the measurement interval
is widened to three years. However, it drops when the measurement
interval is four years. The explanatory power of the model increases
from 7.1 to 23.6%. These results are comparable with those presented in
the study by Kothari and Sloan (1992). An additional interpretation of
the findings is that stock prices reflect investors' view that
current earnings provide them with information about future earnings and
future returns considering the firms listed in ASE, i.e. the market
responds quickly to new information and anticipates earnings. Within
this interpretation, the results suggest that the Athens stock exchange
market demonstrates some features of semi-strong efficiency, and stock
prices incorporate all publicly available information.
The present study can be further extended for examining the
relationship between returns and accounting earnings in a longer period,
as well as other econometric methods can be used for improving the final
results. For further analysis, it would be motivating to take control
over permanent and transitory components of earnings and evaluate the
information content of other financial statements additionally to
earnings in order to acknowledge value-relevant events not recognized in
the earnings of the period under consideration. Moreover, the
examination of the effect of losses (negative earnings) in the
earnings-returns relation would be a prolific path for future research.
Caption: Fig. 1. Measurement intervals of earnings and returns
doi: 10.3846/16111699.2012.758168
Source: Kothari, Sloan (1992: 148)
References
Amir, E.; Lev, B. 1996. Value-relevance of nonfinancial
information: the wireless communications industry, Journal of Accounting
and Economics 22(1/3): 3-30.
http://dx.doi.org/10.1016/S0165-4101(96)00430-2
Ball, R.; Brown, P. 1968. An empirical evaluation of accounting
income numbers, Journal of Accounting Research 6(2): 159-178.
http://dx.doi.org/10.2307/2490232
Beaver, W.; Lambert, R.; Morse, D. 1980. The information content of
security prices, Journal of Accounting and Economics 2(1): 3-28.
http://dx.doi.org/10.1016/0165-4101(80)90013-0
Brown, L.; Griffin, P.; Hagerman, R.; Zmijewski, M. 1987. An
evaluation of alternative proxies for the market's expectation of
earnings, Journal of Accounting and Economics 9(2): 159-193.
http://dx.doi.org/10.1016/0165-4101(87)90004-8
Collins, D. W.; Kothari, S. P. 1989. An analysis of intertemporal
and cross-sectional determinants of earnings response coefficients,
Journal of Accounting and Economics 11(2/3): 143-181.
http://dx.doi.org/10.1016/0165-4101(89)90004-9
Collins, D. W.; Kothari, S. P.; Shanken, J.; Sloan, R. 1994. Lack
of timeliness versus noise as explanations for low contemporaneous
return-earnings association, Journal of Accounting and Economics 18(3):
289-324. http://dx.doi.org/10.1016/0165-4101(94)90024-8
Dimitropoulos, P.; Asteriou, D. 2009. The relationship between
earnings and stock returns: empirical evidence from the Greek capital
market, International Journal of Economics and Finance 1(1): 40-50.
Donnelly, R.; Walker, M. 1995. Share price anticipation of earnings
and the effect of earnings persistence and firm size, Journal of
Business, Finance and Accounting 22(1): 5-18.
http://dx.doi.org/10.1111/j.1468-5957.1995.tb00668.x
Dumontier, P.; Labelle, R. 1998. Accounting earnings and firm
valuation: the French case, The European Accounting Review 7(2):
163-183. http://dx.doi.org/10.1080/096381898336439
Easton, P. D.; Harris, T. S.; Ohlson, J. 1992. Aggregate accounting
earnings can explain most security returns: the case of long return
intervals, Journal of Accounting and Economics 15(2/3): 119-42.
http://dx.doi.org/10.1016/0165-4101(92)90015-T
Easton, P. D.; Harris, T. S. 1991. Earnings as an explanatory
variable for returns, Journal of Accounting Research 29(1): 19-36.
http://dx.doi.org/10.2307/2491026
Easton, P.; Zmijewski, M. 1989. Cross-sectional variation in the
stock market response to accounting earnings announcements, Journal of
Accounting and Economics 11(2/3): 49-36.
Hayn, C. 1995. The information content of losses, Journal of
Accounting and Economics 20(2): 125-53.
http://dx.doi.org/10.1016/0165-4101(95)00397-2
Jarmalaite, N. 2002. The relationship between accounting numbers
and returns in the Baltic Stock Markets, CERT Discussion paper No
2002/06. Edinburgh: Heriot-Watt University.
Jermakowicz, E. K.; Gornik-Tomaszewski, S. 1998. Information
content of earnings in the Emerging Capital Market: evidence from Warsaw
Stock Exchange, Multinational Finance Journal 2(4): 245-267.
Jindrichovska, I. 2001. The relationship between accounting numbers
and returns: some empirical evidence from the Emerging Market of Czech
Republic, The European Accounting Review 10(1): 107-131.
Kallunki, J.; Martikainen, T. 1997. The lead-lag structure of stock
returns and accounting earnings: implications to the returns-earnings
relation in Finland, International Review of Financial Analysis 6(1):
37-47. http://dx.doi.org/10.1016/S1057-5219(97)90018-1
Kormendi, R.; Lipe, R. 1987. Earnings innovations, earnings
persistence and stock returns, Journal of Business 60(3): 323-345.
http://dx.doi.org/10.1086/296400
Kothari, S. P. 1992. Price-earnings regressions in the presence of
prices leading earnings: earnings level versus change specifications and
alternative deflators, Journal of Accounting and Economics 15(2/3):
173-302. http://dx.doi.org/10.1016/0165-4101(92)90017-V
Kothari, S. P.; Sloan, R. 1992. Information in prices about future
earnings: implications for earnings response coefficients, Journal of
Accounting and Economics 15(2/3): 143-171.
http://dx.doi.org/10.1016/0165-4101(92)90016-U
Kothari, S. P.; Zimmerman, J. L. 1995. Price and return models,
Journal of Accounting and Economics 20(2): 155-192.
http://dx.doi.org/10.1016/0165-4101(95)00399-4
Kousenidis, D. 2005. Earnings-returns relation in Greece: some
evidence on the size effect and on the life-cycle hypothesis, Managerial
Finance 31(2): 24-54. http://dx.doi.org/10.1108/03074350510769488
Liu, J.; Thomas, J. 2000. Stock returns and accounting earnings,
Journal of Accounting Research 38(1): 71-102.
http://dx.doi.org/10.2307/2672923
Maditinos, D.; Sevic, Z.; Chatzoglou, P.; Theriou, G. 2007.
Earnings as an explanatory tool in explaining stock market returns and
the use of Easton and Harris (1991) model: the case of Greece, in
International Scientific Conference, Contemporary Challenges of Theory
and Practice in Economics. Belgrade, 26-29 September, 2007, Proceedings.
Belgrade: Faculty of Economics of the University of Belgrade Publishing
Centre, Cugura Print, 2007, 93-106. ISBN 978-86-403-0843-4.
Markowitz, H. M. 1959. Portfolio Selection: Efficient
Diversification of Investments. New York: Wiley.
Markowitz, H. M. 1952. Portfolio selection, Journal of Finance
7(1): 77-91.
Martikainen, T.; Kallunki, J.; Perttunen, J. 1997. Finnish earnings
response coefficients: the information content of losses, The European
Accounting Review 6(1): 69-81. http://dx.doi.org/10.1080/096381897336872
Myring, M., Shortridge, R. T.; Alst, L. V. 2003. The relationship
between returns and unexpected earnings: a comparison of Australia and
the United States, Journal of Business and Economics Research 1(8):
71-79.
Ohlson, J. 1995. Earnings, book values, and dividends in equity
valuation, Contemporary Accounting Research 11(2): 661-687.
http://dx.doi.org/10.1111/j.1911-3846.1995.tb00461.x
Ohlson, J. A.; Shroff, P. K. 1992. Changes versus levels in
earnings as explanatory variables for returns: some theoretical
considerations, Journal of Accounting Research 30(2): 210-226.
http://dx.doi.org/10.2307/2491124
Papadaki, A.; Siougle, G. 2007. Value relevance of price, earnings
and book values in the Athens Stock Exchange, Managerial Finance 33(5):
309-320. http://dx.doi.org/10.1108/03074350710739597
Stankeviciene, J.; Gembickaja, N. 2012. Market behaviour: case
studies of NASDAQ OMX Baltic, Business, Management and Education 10(1):
110-127.
Suwardi, E. 2009. The dynamic relationship between accounting
numbers and share prices on the Jakarta Stock Exchange, International
Review of Business Research Papers 5(5): 16-24.
Dimitrios I. Maditinos (1), Zeljko Sevic (2), Jelena Stankeviciene
(3), Nikolaos Karakoltsidis (4)
(1,4) Kavala Institute of Technology, School of Business and
Economics, Department of Business Administration, Ag. Loukas, Kavala,
65404 Greece
(2) Glasgow Caledonian University, Glasgow School for Business and
Society, Cowcaddens Road, Glasgow G4 0BA, United Kingdom
(3) Department of Finance Engineering, Vilnius Gediminas Technical
University, Sauletekio al. 11, LT-10223 Vilnius, Lithuania
E-mails: (1) dmadi@teikav.edu.gr; (2) zeljko.sevic@gcal.ac.uk; (3)
jelena.stankeviciene@vgtu.lt (corresponding author); (4)
karakoltsidis@hotmail.com
Received 18 December 2010; accepted 15 February 2012
Dimitrios I. MADITINOS is an Assoc. Professor of Information
Technology, Finance and Financial Modelling at Kavala Institute of
Technology, School of Business and Economics (Greece). His research
interests are in financial modelling, performance measurement systems,
investors' behaviour, financial information systems, ERPs,
electronic commerce and e-business.
Zeljko SEVIC is a Professor at Glasgow Caledonian University, Dean
of Caledonian Business School (UK). His current research interests are
in the public sector reform in Japan and Vietnam (on-going projects),
and accounting reform in the public sector with the particular emphasis
on the use of management accounting in public sector restructuring, and
the design of effective financial/fiscal information system.
Jelena STANKEVICIENE is an Assoc. Professor in the Department of
Finance Engineering at Vilnius Gediminas Technical University
(Lithuania). Her main research topics include assets and liability
management, regulation of financial institution, financial management
for value creation, value engineering.
Nikolaos KARAKOLTSIDIS hold an MSc in Finance and Financial
Information Systems from the Business School, Greenwich University,
London, UK. His research interests are in financial modelling,
performance measurement systems, investors' behaviour, financial
information systems, ERPs, electronic commerce and e-business.
Table 1. Descriptive statistics
Panel: 1 Mean St. dev. Median Min. Max. N
Cumulative
earnings yield
Length of the
period
1 year 0.128 0.310 0.048 -1.640 5.836 2.166
2 years 0.253 0.547 0.108 -2.849 6.845 1.910
3 years 0.377 0.731 0.172 -3.493 7.793 1.675
4 years 0.017 0.105 0.009 -1.764 3.422 1.441
Panel: 2 Non- Mean St. dev. Median Min. Max. N
cumulative
earnings yield
Length of the
period
1 year 0.128 0.310 0.048 -1.640 5.836 2.166
2 years 0.131 0.430 0.055 -4.747 7.083 1.910
3 years 0.129 0.433 0.052 -4.093 9.437 1.675
4 years 0.114 0.344 0.044 -3.993 4.239 1.441
Panel: 3 Price Mean St. dev. Median Min. Max. N
relatives,
including
dividend yield
Length of the
period
1 year 2.061 0.904 0.008 22.941 2.166
2 years 1.803 3.870 0.839 0.009 55.631 1.910
3 years 1.389 2.485 0.813 0.007 52.487 1.675
4 years 1.193 0.715 0.016 0.016 18.481 1.441
Notes: Earnings yield ([X.sub.it-[tau],t]/[P.sub.it-[tau]]) is annual
earnings per share before extraordinary non-recurring items such as
discontinued operations and special items over years t-[tau] to t
divided by the price at the beginning of the return measurement
interval. Price relative ([P.sub.it]/[P.sub.it-[tau]]) is one plus
buy-and-hold return inclusive of dividends over years t-[tau] to t.
All earnings per share numbers and prices are adjusted for stock
splits. N is the number of observations.
Table 2. Regressions of price relatives to earnings deflated by price
[P.sub.it]/[P.sub.it-[tau]] = [a.sub.i]
+ [[gamma].sub.1i] [X.sub.it]/
[P.sub.it-[tau]] + [[epsilon].sub.it]
Yearly Model
[tau] = 1 [tau] = 2 [tau] = 3 [tau] = 4
All Intercept 0.317 0.674 0.695 0.605
Years
t (8.75) (14.105) (22.867) (11.638)
Sig. 0.000 0.000 0.000 0.000
ERC 0.139 0.266 0.317 0.321
t (12.914) (19.003) (22.867) (22.136)
Sig. 0.000 0.000 0.000 0.000
[R.sup.2] 0.072 0.16 0.238 0.255
Adj 0.071 0.159 0.238 0.254
[R.sup.2]
N 2,166 1,910 1,675 1,441
Notes: The interval of earnings measurement ranges from one to four
years. Price relatives are contemporaneous for one year and periods
that include leading years. Variables have been measured using OLS
regression analysis of data for the period 1998-2008. Sample sizes
vary from 2166 to 1441 firm year observations (t-statistic is
significant at the level of 5%). N is the number of observations in
the regression.
Table 3. The regression of price relatives to earnings deflated by
price
[P.sub.it]/[P.sub.it-[tau]] = [a.sub.i] +
[[gamma].sub.1i] [X.sub.it]/[P.sub.it-[tau]] +
[e.sub.it]
Yearly Model
[tau] = 1 [tau] = 2 [tau] = 3 [tau] = 4
1 2 3 4 5 6
1998 Intercept 0.726
t (7.688)
Sig. 0.000
ERC 0.085
t (2.44)
Sig. 0.016
R2 0.052
Adj R2 0.043
N 110
1999 Intercept 2.448 2.798
t (18.078) (24.068)
Sig. 0.000 0.000
ERC 0.21 0.196
t (5.23) (4.894)
Sig. 0.000 0.000
[R.sup.2] 0.178 0.182
Adj [R.sup.2] 0.172 0.174
N 128 110
2000 Intercept -1.07 1.262 1.537
t (-20.28) (11.174) (15.441)
Sig. 0.000 0.000 0.000
ERC -0.026 0.224 0.17
t (-2.003) (5.823) (4.224)
Sig. 0.047 0.000 0.000
[R.sup.2] 0.027 0.213 0.142
Adj [R.sup.2] 0.02 0.207 0.134
N 149 128 110
2001 Intercept 0.71 -0.808 1.096 1.437
t (0.875) (-5.095) (7.661) (12.845)
Sig. 0.383 0.000 0.000 0.000
ERC 0.119 0.144 0.261 0.243
t (5.591) (4.265) (5.659) (5.791)
Sig. 0.000 0.000 0.000 0.000
[R.sup.2] 0.142 0.104 0.203 0.235
Adj [R.sup.2] 0.137 0.103 0.196 0.228
N 191 151 128 110
2002 Intercept -0.22 -0.138 -1.006 0.611
t (-2.506) (-1.169) (5.559) (4.161)
Sig. 0.013 0.244 0.000 0.000
ERC 0.105 0.216 0.219 0.28
t (4.128) (7.014) (5.936) (6.409)
Sig. 0.000 0.000 0.000 0.000
[R.sup.2] 0.079 0.207 0.191 0.246
Adj [R.sup.2] 0.074 0.202 0.186 0.24
N 202 191 151 128
2003 Intercept 0.236 -0.048 -0.682 -1.1014
t (4.509) (-0.541) (-5.49) (-5.96)
Sig. 0.000 0.589 0.000 0.000
ERC 0.37 0.108 0.023 0.176
t (2.242) (4.531) (0.684) (5.265)
Sig. 0.026 0.000 0.495 0.000
[R.sup.2] 0.023 0.093 0.02 0.157
Adj [R.sup.2] 0.019 0.089 -0.03 0.151
N 213 202 191 151
2004 Intercept -0.175 0.096 9.219 -0.334
t (-2.487) (1.12) (-1.791) (-2.39)
Sig. 0.014 0.264 0.075 0.018
ERC 0.051 0.105 0.152 0.201
t (2.41) (3.946) (4.617) (5.891)
Sig. 0.017 0.000 0.000 0.000
[R.sup.2] 0.026 0.069 0.096 0.155
Adj [R.sup.2] 0.021 0.064 0.092 0.151
N 221 213 202 191
2005 Intercept 0.046 0.259 0.534 -0.163
t (0.664) (2.604) (4.895) (-1.108)
Sig. 0.507 0.010 0.000 0.269
ERC -0.014 0.17 0.23 0.147
t (-0.592) (5.653) (6.706) (3.64)
Sig. 0.504 0.000 0.000 0.000
[R.sup.2] 0.002 0.128 0.176 0.062
Adj [R.sup.2] -0.003 0.124 0.172 0.058
N 231 221 213 201
2006 Intercept 0.547 0.732 0.689 0.954
t (9.385) (8.973) (6.685) (8.625)
Sig. 0.000 0.000 0.000 0.000
ERC 0.102 0.144 0.222 0.278
t (5.368) (5.316) (7.247) (8.13)
Sig. 0.000 0.000 0.000 0.000
[R.sup.2] 0.11 0.11 0.194 0.239
Adj [R.sup.2] 0.106 0.106 0.19 0.236
N 235 231 221 212
2007 Intercept 0.239 0.759 0.992 0.926
t (3.5) (9.329) (9.793) (7.813)
Sig. 0.001 0.000 0.000 0.000
ERC 0.048 0.148 0.212 0.281
t (2.207) (5.342) (5.989) (7.556)
Sig. 0.028 0.000 0.000 0.000
[R.sup.2] 0.02 0.11 0.135 0.207
Adj [R.sup.2] 0.016 0.106 0.132 0.204
N 235 234 231 221
2008 Intercept -0.975 0.732 0.687 -0.333
t (-11.027) (8.994) (6.696) (-3.618)
Sig. 0.000 0.000 0.000 0.000
ERC -0.032 0.144 0.221 0.042
t (-1.281) (5.338) (7.271) (-3.618)
Sig. 0.217 0.000 0.000 0.177
[R.sup.2] 0.007 0.01 0.022 0.008
Adj [R.sup.2] 0.003 0.006 0.018 0.004
N 235 228 228 227
Notes: The interval of earnings measurement ranges from one to four
years. Price relatives are contemporaneous for one year and periods
that include leading years. Variables have been measured using OLS
regression analysis of data for the period 1998-2008 (t-statistic is
significant at the level of 5%). N is the number of observations in
the regression.
Table 4. The regression of price relatives to cumulative earnings
deflated by price
[P.sub.it]/[P.sub.it-[tau]] =
[a.sub.i] + [[gamma].sub.1i]
[X.sub.it-[tau], t]/[P.sub.it-[tau]]
+ [e.sub.it]
Cumulative Yearly Model
[tau] = 1 [tau] = 2 [tau] = 3 [tau] = 4
All Intercept 0.317 0.027 0.264 0.027
Years
t (8.75) (11.554) (7.727) (0.821)
Sig. 0.000 0.000 0.000 0.412
ERC 0.139 0.275 0.328 0.321
t (12.914) (18.160) (21.704) (21.111)
Sig. 0.000 0.000 0.000 0.000
[R.sup.2] 0.072 0.237 0.220 0.237
Adj [R.sup.2] 0.071 0.147 0.220 0.236
N 2,166 1,910 1,675 1,441
Notes: The interval of earnings measurement ranges from one to
four years. Price relatives are contemporaneous for one year and
periods that include leading years. Data on earnings are
cumulative. Variables have been measured using OLS regression
analysis of data for the period 1998-2008. Sample sizes vary from
2,166 to 1,441 firm year observations (t-statistic is significant
at the level of 5%). N is the number of observations in the
regression.
Table 5. The regression of price relatives to cumulative earnings
deflated by price
[P.sub.it]/[P.sub.it-[tau]] = [a.sub.i] +
[[gamma].sub.1i] [X.sub.it-[tau], t]/
[P.sub.it-T] + [e.sub.it]
Cumulative Yearly Model
[tau] = 1 [tau] = 2 [tau] = 3 [tau] = 4
1 2 3 4 5 6
1998 Intercept 0.726
t (7.688)
Sig. 0.000
ERC 0.085
t (2.44)
Sig. 0.016
[R.sup.2] 0.052
Adj [R.sup.2] 0.043
N 110
1999 Intercept 2.448 2.683
t (18.078) (25.451)
Sig. 0.000 0.000
ERC 0.21 0.239
t (5.23) (4.383)
Sig. 0.000 0.000
[R.sup.2] 0.178 0.151
Adj [R.sup.2] 0.172 0.143
N 128 110
2000 Intercept -1.07 1.177 1.391
t (-20.28) (12.847) (20.096)
Sig. 0.000 0.000 0.000
ERC -0.026 0.270 0.233
t (-2.003) (6.595) (5.102)
Sig. 0.047 0.000 0.000
[R.sup.2] 0.027 0.257 0.194
Adj [R.sup.2] 0.02 0.251 0.187
N 149 128 110
2001 Intercept 0.71 -1.480 0.591 0.894
t (0.875) (-15.166) (7.260) (12.587)
Sig. 0.383 0.000 0.000 0.000
ERC 0.119 -0.080 0.170 0.013
t (5.591) (-0.338) (5.080) (1.540)
Sig. 0.000 0.736 0.000 0.126
[R.sup.2] 0.142 0.001 0.172 0.022
Adj [R.sup.2] 0.137 -0.006 0.163 0.012
N 191 151 128 110
2002 Intercept -0.22 -0.554 -1.576 -0.031
t (-2.506) (-7.154) (-14.786) (-0.382)
Sig. 0.013 0.000 0.000 0.703
ERC 0.105 0.140 0.149 0.199
t (4.128) (5.678) (5.035) (5.605)
Sig. 0.000 0.000 0.000 0.000
[R.sup.2] 0.079 0.146 0.145 0.200
Adj [R.sup.2] 0.074 0.141 0.140 0.193
N 202 191 151 128
2003 Intercept 0.236 -0.080 -0.396 -1.389
t (4.509) (-0.977) (-5.347) (-14.295)
Sig. 0.000 0.330 0.000 0.000
ERC 0.037 0.137 0.168 0.170
t (2.242) (4.546) (6.457) (5.903)
Sig. 0.026 0.000 0.000 0.000
[R.sup.2] 0.023 0.094 0.181 0.190
Adj [R.sup.2] 0.019 0.089 0.176 0.184
N 213 202 191 151
2004 Intercept -0.175 0.030 -0.288 -0.707
t (-2.487) (0.407) (-2.905) (-8.875)
Sig. 0.014 0.685 0.004 0.000
ERC 0.051 0.120 0.228 0.205
t (2.41) (3.805) (5.189) (6.630)
Sig. 0.017 0.000 0.000 0.000
[R.sup.2] 0.026 0.064 0.119 0.189
Adj [R.sup.2] 0.021 0.060 0.114 0.184
N 221 213 202 191
2005 Intercept 0.046 0.083 0.233 -0.158
t (0.664) (0.947) (2.384) (-1.610)
Sig. 0.507 0.345 0.005 0.109
ERC -0.014 0.162 0.253 0.315
t (-0.592) (4.34)2 (5.577) (6.179)
Sig. 0.504 0.000 0.000 0.002
[R.sup.2] 0.002 0.079 0.128 0.161
Adj [R.sup.2] -0.003 0.075 0.124 0.157
N 231 221 213 201
2006 Intercept 0.547 0.683 0.487 0.544
t (9.385) (10.198) (6.219) (7.424)
Sig. 0.000 0.000 0.000 0.000
ERC 0.102 0.186 0.295 0.346
t (5.368) (6.028) (7.367) (7.562)
Sig. 0.000 0.000 0.000 0.000
[R.sup.2] 0.11 0.137 0.199 0.213
Adj [R.sup.2] 0.106 0.133 0.195 0.210
N 235 231 221 212
2007 Intercept 0.239 0.709 0.811 0.579
t (3.5) (10.760) (11.484) (7.485)
Sig. 0.001 0.000 0.000 0.000
ERC 0.048 0.189 0.276 0.379
t (2.207) (6.147) (6.468) (7.916)
Sig. 0.217 0.000 0.000 0.000
[R.sup.2] 0.02 0.140 0.157 0.222
Adj [R.sup.2] 0.016 0.136 0.153 0.219
N 235 235 231 221
2008 Intercept -0.983 -0.704 0.403 -0.365
t (-11.106) (-8.618) (-8.793) (-5.932)
Sig. 0.000 0.000 0.000 0.000
ERC -0.033 0.045 0.098 0.094
t (-1.281) (1.000) (2.611) (2.365)
Sig. 0.000 0.187 0.010 0.019
[R.sup.2] 0.007 0.008 0.171 0.024
Adj [R.sup.2] 0.003 0.003 0.025 0.020
N 235 228 228 227
Notes: The interval of earnings measurement ranges from one to four
ears. Price relatives are contemporaneous for one year and periods
that include leading years. Variables have been measured using OLS
regression analysis of data for the period 1998-2008 (t-statistic is
significant at the level of 5%). N is the number of observations in
the regression.
