1 Chapter 1 INTRODUCTION 1. 1 General Background From the past decades, the financial market has been suffering from the unforeseen and sudden economic turbulences that have been directly or indirectly contributing for stock returns movements. Identifying the factors affecting stock returns is not an easy task for the financial economists, academicians and practitioners. To grasp some ideas through the systematic procedure, the financial community felt the need of separate discipline so that the new discipline can solely deals with the management of financial assets. The investment management and the portfolio theories are the outcomes of such efforts. The evolution of investment management and the portfolio theory have long history. The development of investment management can be traced chronologically through three different phases (Francis, 1986). The first phase could be characterized as the speculative phase before 1929. During the 1930s investment management entered in its second phase, a phase of professionalism. Then, the investment industry began the process of upgrading its ethics, establishing standard practices and generating a good public image. As a result, the investment markets became safer places and the ordinary people also began to invest. Investors began to analyze the securities seriously before undertaking investments. Then, the investment community entered into its third phase, the scientific phase after Markowitz’s study in 1952.
311
Embed
Market information and stock returns the nepalese evidence
This is a Nepalese Stock Market research on the market information and its effects on stock price. More specifically, this study gauge the political effect, media effect, news coverage effect, determine the investors' priority prior to making investment decisions and finally and most importantly, the study provides the evidences that how long of historical data base are useful for investment decision making. I hope every one enjoy the research work.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
Chapter 1
INTRODUCTION1. 1 General Background
From the past decades, the financial market has been suffering from the unforeseen and
sudden economic turbulences that have been directly or indirectly contributing for stock
returns movements. Identifying the factors affecting stock returns is not an easy task for
the financial economists, academicians and practitioners. To grasp some ideas through the
systematic procedure, the financial community felt the need of separate discipline so that
the new discipline can solely deals with the management of financial assets. The
investment management and the portfolio theories are the outcomes of such efforts. The
evolution of investment management and the portfolio theory have long history. The
development of investment management can be traced chronologically through three
different phases (Francis, 1986). The first phase could be characterized as the speculative
phase before 1929. During the 1930s investment management entered in its second phase,
a phase of professionalism. Then, the investment industry began the process of upgrading
its ethics, establishing standard practices and generating a good public image. As a result,
the investment markets became safer places and the ordinary people also began to invest.
Investors began to analyze the securities seriously before undertaking investments. Then,
the investment community entered into its third phase, the scientific phase after
Markowitz’s study in 1952.
Markowitz (1952) which is a single-period model and attempted to quantify the risk and it
showed that the risk in investment could be reduced through proper diversification of
investment which required the creation of a portfolio. The Markowitz study was extended
to CAPM in 1960s by Sharpe (1964), Lintner (1965) and Black (1972). The CAPM
explains the overall market performance that determines the stock returns. Then, the
assets valuation models became the most popular area of study in Finance in developed,
developing and the transitional economies. In other words, the history of development of
the portfolio theories and its practices enter into the professionalism and scientific phase.
The empirical evidences of Stattman (1980), Chan, et.al (1991), Brav, et.al (2000), Daniel
and Titman (2006) among others documented the book-to-market equity effects on stock
returns; earnings-to-price effects by Basu (1977), earning effects by Jafee, et.al (1989),
Fama and French (1995) and La Porta (1996) among others; Banz (1981), Vassalou and
Xing (2004) and Fama and French (2008) depicted the size effects, similarly, cash flows
2
effects by Berk, et.al (1999) and Vuolteenaho (2002) among others are the major studies
that documented the firm specific accounting variables are the major sources of stock
returns changes. Whereas in the later period, more focus was given towards the
behavioral aspects like investors’ characteristics and behavioral issues and market
behavior, news effects, media effects, etc. In sum, the recent focus has shifted towards the
intangibles rather than the fundamental effects on stock returns. The studies on human
psychology and behavioral issues, Einhorn, et al. (1978) documented that people have
great confidence in their fallible judgment. Similarly, Einhorn (1980) conformed that the
overconfidence in judgment showed that the contribution of behavioral factors in stock
The review covers the major studies on respective variables during 1981 to 2008.
i) Review of major studies on book-to-market effects
The book-to-market effects on stock returns from the period 1991 to 2006 have been
presented in the Table 2.1 below:
Table 2.1: Review of major studies on book-to-market effectsStudy Major findingsChan, et.al (1991) The book to market ratio and cash flow yield has the most significant positive
impact on expected returns.
Davis (1994) Book-to-market equity, earnings yield, and cash flow yield have significant explanatory power with respect to the cross-section of realized stock returns and, there was a strong January seasonal in the explanatory power of book-to-market equity, earning yield and cash flow yield.
Brav, et.al (2000) Underperformance is concentrated primarily in small issuing firms with low book-to-market ratios.
Daniel and Titman (2006)
Book-to-market equity ratio, a good proxy for intangible return forecasts returns. A composite equity issuance measure, also as an intangible information independently forecasts the future returns.
The study analyzed the cross-sectional differences of Japanese stocks returns to the
underlying behavior of the variables: earnings yield, size, book-to-market ratio, and cash
flow yield, Chan, et.al (1991). Seemingly Unrelated Regression (SUR) model and Fama-
Mac-Beth (1973) methodology are applied on comprehensive, high-quality monthly data
set of stocks listed on the Tokyo Stock Exchange (TSE) that extends from 1971 to 1988.
The sample includes both manufacturing and nonmanufacturing firms, companies from
both sections of the Tokyo Stock Exchange, and also delisted securities. The findings
revealed the univariate analysis of stock returns and fundamental variables indicated that
high earnings yield stocks outperform low earnings yield stocks; small stocks achieved
substantially higher returns than large stocks; the firms with large positive book-to-
market equity ratio earned high premium than firms with low; and positive book-to-
market equity. Further, cash flows yield is found to have positive relation with stock
returns. However regression analysis produced striking results: the earning yield effect
was not significant across the different regressions models and it was not even significant
when earning yield was the only independent variables. Firm size, in general, is
significant with an unexpected sign meaning that large companies in Japan tend to
outperform small companies. The performance of book-to-market equity is statistically
and economically the most important among the four variables investigated. Although,
the study confirmed the existence of size effect after adjusting for market risk and other
17
fundamental variables, the statistical significant of the size variable is sensitive to the
specification of the model. Of the four variables investigated, though, it is hardest to
disentangle the effect of earnings yield variable. In sum, the book to market ratio and cash
flow yield has the most significant positive impact on expected returns.
The 100 firms listed in the Moody’s Industrial Manual which is free from the
survivorship bias and four fundamental variables: book-to-market equity, cash flow yield,
earning yield and historical sales growth as primary focus of the study (Davis, 1994).
Stock returns, stock prices and market values of equity were derived from the CRSP
monthly file. The study uses the Fama-MacBeth (1973) cross-sectional regression model
to determine the explanatory power of realized returns from 1940 to the early 1960s, the
pre-COMPUSTAT era. The findings of the study includes: significant relationship
between book-to-market equity and subsequent returns, cash flow yield has explanatory
power with respect to subsequent realized returns when book-to-market equity and
historical sales growth are held constant, earning yield has also explanatory power to
predict subsequent returns, insignificant explanatory power for beta to predict returns,
weak relationship between sales growth and returns, and there is significance of log book-
to-market, earning to price, and cash flow to price with returns mostly in January. Thus,
the study concluded that book-to-market equity, earnings yield, and cash flow yield have
significant explanatory power with respect to the cross-section of realized stock returns
and there was a strong January seasonal in the explanatory power of book-to-market
exists is the major focus of the study. Sample of initial public offering (IPO) and seasoned
equity offering (SEO) of the firms from 1975 to 1992 derived from CRSP for NYSE,
ASE and NASDAQ. The sample included 4526 offerings made by 2772 firms. The study
found that underperformance is concentrated primarily in small issuing firms with low
book-to-market ratios. SEO firms that underperform these standard benchmarks have
time series returns that covary with factor returns constructed from non-issuing firms. The
study concluded that the stock returns following equity issues reflect a more pervasive
return pattern in broader set of publicly traded companies.
Book-to-market equity ratio forecasts stock returns because it is a good proxy for
intangible returns. Further, composite equity issuance measure, which is related to
intangible returns, independently forecasts returns (Daniel and Titman, 2006). The book-
18
to-market effect is often interpreted as evidence of high expected returns on stocks of
distressed firms with poor past performance. The study also found that while a stock’s
future return is unrelated to the firm’s past accounting-based performance, it is strongly
negatively related to the intangible return, the component of its past return that is
orthogonal to the firm’s past performance. Other findings of the study are: stock returns
over a relatively long horizon (5 years) should be closely linked to concurrent
fundamental performance; there is a strong positive relation between intangible returns
and future fundamental performance measures i.e. a firm’s intangible returns reflects, at
least partial information to its future growth prospects, there is no evidence of any link
between past tangible information and future return; there is strong negative relation
between past tangible returns and future returns; future returns are unrelated to internally
funded growth in sales; future returns are strongly negatively associated with growth that
is financed by the share issuance; composite share issuance variable is significantly
negatively related to future returns; the strong intangible return and issuance effects that
cannot explain the existence of mispricing; low book-to-market firms have both higher
future accounting growth rates and lower future returns; negative correlation between the
lagged book-to-market ratio and book-return; and the composite share issuance measure
is strongly negatively related to future returns.
ii) Review of major studies on cash flow and earnings effects
The cash flow is considered as the fundamental variable and the variation in cash flow
might be the causes of changes in the stock returns. The major previous studies on cash
flow from 1999 to 2002 have been presented in Table 2.2 as follows:
An analysis on optimal investment, growth options and security returns is conducted by
Berk, et.al (1999). The interest of the study is the individual firm. The random evolution
of the firm’s collection of projects determines how its risk and return change over time. In
the study, the partial equilibrium model gives the tractability to focus on the dynamics for
the relative risks of individual firms. The study found that as a consequence of optimal
investment choices, a firm’s assets and growth options change in predictable ways. In the
study, the dynamic model imparts predictability to changes in a firm’s systematic risk,
and its expected returns. Simulations showed that the model
Table 2.2: Review of major studies on cash flow and earnings effectsStudy Major findingsBerk, et.al (1999) The valuation of the cash flows that result from the investment decision making by
the individual firms, along with the firm’s opinions to grow in the future, leads to dynamics for conditional expected returns.
19
Vuolteenaho (2002)
Firm-level stock returns are mainly driven by cash-flow news.
Jafee, et.al (1989) A significant relation between returns and earnings only in the month of January and, the size effect was negative only in January.
Fama and French (1995)
There are market, size, and BE/ME factors in earnings like those in returns.
La Porta (1996) Earnings growth is the only variable with the significant explanatory power in explaining stock returns.
simultaneously reproduces: the time-series relation between the book-to-market ratio and
asset returns; the cross-sectional relation between book-to-market ratio, market value, and
returns, contrarian effects at short horizons; momentum effects at longer horizons and the
inverse relation between interest rates and the market risk premium. The study simulated
20,000 months of data for 50 firms and restrict the attention to firms that have reached a
steady state distribution for the number of ongoing projects by dropping the first 200
observations. In addition to dynamic and simulation models, FM regression models,
varying types of frequency distributions are used for the analysis. The findings of the
study concluded that the valuation of the cash flows that result from the investment
decision making by the individual firms, along with the firm’s opinions to grow in the
future, leads to dynamics for conditional expected returns. The model of expected returns
in the study helps explain a number of the important features of the cross-sectional and
time-series behavior of stock returns, and the biases that might be induced by the model
that ignores these dynamics. On the other hand, the simulation results showed that the
model can reproduce simultaneously several important cross-sectional and time-series
behaviors that studies have documented for stock returns, including the explanatory
power of book-to-market value, and interest rates, and the success of contrarian and
momentum strategies at different horizons.
Vuolteenaho (2002) conducted a study on firm-level returns, where author use a vector
autoregressive model (VAR) to decompose an individual firm’s stock returns into two
components: changes in cash-flow expectations i.e. cash-flow news and changes in
discount rates i.e. expected-return news. By definition, a firm’s stock returns are driven
by shocks to expected cash flows (cash-flow news) and/or shocks to discount rates
(expected-return news). Substantial studies have been done to measure the relative
importance of cash-flow and expected-return news for aggregate portfolio returns, but
virtually no evidence on the relative importance of these components at the firm level.
The basic data for the study derived from the CRSP-COMPUSTAT intersection, from
1954 to 1996. CRSP monthly stock file contains the data of monthly prices, shares
outstanding, dividends, and returns for NYSE, AMEX, and NASDAQ stocks,
20
COMPUSTAT contains the relevant accounting information for the most publicly traded
U.S. stocks and the study, in addition, used rolled-over one month Treasury-bill returns as
risk-free rate. Based on the VAR and Campbell’s (1991) return-decomposition
framework enable the study to decompose the firm-level stock returns into cash-flow and
expected-return news and to estimate how important these two sources of stock variation
are for an individual firm. In addition, the study measure whether positive cash-flow news
is typically associated with an increase or decrease in expected returns. The findings of
the study includes - the information about future cash flows is the dominant factor driving
firm-level stock returns, cash-flow news is positively correlated with expected returns for
a typical stock. Finally, it is appeared that while cash-flow information is largely firm
specific, expected-return information is predominantly driven by systematic,
macroeconomic components. In sum, VAR yields three main results. First, firm-level
stock returns are mainly driven by cash-flow news. For a typical stock, the variance of
cash-flow news is more than twice that of the expected-return news. Second, the expected
returns and cash flows are positively correlated for a typical small stock. Third, expected-
return-news series are highly correlated across firms, while cash-flow news can largely be
diversified away in aggregate portfolios.
The study uses the CRSP monthly stock return data for relatively a longer period from
1951 to 1986 and from the “back data” versions from 1950-1966 periods. Jafee, et.al
(1989) evaluated the relation between size and earnings yield effects on stock returns.
Over the entire period, the study reported a significant relationship earnings and stock
returns only in the month of January, while it is observed a significant relation during all
months of the sub-period 1969-1986. Conversely, the size effect is found significantly
negative only in January in the overall period and in both sub-periods.
Fama and French (1995) analyzed whether the behavior of stock prices in relation to size
and book-to market-equity (BE/ME) reflects the behavior of earnings. The study focused
on six portfolios, formed yearly from a simple sort of firms in to two group on market
equity and another simple sort into three groups on book-to-market equity. Further, it is
study attempted to provide an economic foundation for empirical relations between
average stock returns and size, and average stock returns and book-to-market equity
observed in Fama and French (1992). Consistent with rational pricing, high BE/ME
prices forecast the reversion of earnings growth observed after firms are ranked on size
and BE/ME. The evidence that size and book to market equity proxy for sensitivity to risk
21
factors in returns is consistent with a rational pricing story for the role of size and BE/ME
in average returns. Specifically, the analysis of whether the behavior of stock prices, in
relation to size and book-to-market equity, is consistent with the behavior of earnings. In
a nutshell, low BE/ME, a high stock price relative to book value, is typical of firms with
high average returns on capital (growth stocks), whereas high BE/ME is typical of firms
that are relatively distressed. Size is also related to profitability, controlling for BE/ME,
small stocks tend to have lower earnings on book equity than do big stocks. The tests
center on six portfolios formed on ranked values of size and BE/ME for individual stocks
i.e. profitability, earnings, profitability in chronological time, earnings/price ratios,
earnings growth rates, and stock returns. Then, the overall analysis examines the links
between returns and these common factors in earnings and established that the level of
earnings is related to size and BE/ME. The study is based on the data from 1963 to 1992
of NYSE, AMEX and NADSAQ. Information was abstracted from the CRSP. Groups are
formed based on the breakpoints for the bottom 30 percent (Low), middle 40 percent
(Medium), and top 30 percent (High) of the ranked values of BE/ME for NYSE stocks
and do not consider the negative BE firms. Thus, the overall relationship of variables
among the portfolios, analysis of regression results suggest that there are market, size, and
BE/ME factors in earnings like those in returns.
Further, La Porta (1996) examined whether investors make the systematic mistakes that
are consistence with the errors in expectation hypothesis when growth in earnings. The
study employed CRSP monthly returns files of the listed companies of NYSE and
AMEX. Annual portfolio returns are constructed by compounding monthly returns. The
regression results reported that earnings growth as the only variable with the significant
explanatory power. The study revealed that the earnings growth is the only significant
variable in multivariate regression when it is combined with size, book-to-market equity,
and cash flow to price ratio. The regression results confirmed the role of the expected rate
of earnings growth in explaining stock returns. The findings are based on multivariate
regression models which reported the negative relation of expected returns with book-to-
market equity, size and earnings growth and positive relation with cash-flow yield. When
stock were sorted by expected growth rate in earnings, it is shown that low earnings
growth stock beat high earnings growth stock by twenty percentage points. The study
further documented that there is no evidence that low earnings growth stocks are more
risky than high earnings growth stocks. When portfolios were formed on the basis of
22
expected growth rate in earnings, the results indicated that low earnings growth stocks
have significantly lower standard deviations and betas than high earning growth stocks.
iii) Review of major studies on size effects
The size is defined as the market value of common stock outstanding and it is also called as fundamental variable for stock returns. The major studies including some seminal works have been presented in Table 2.3 below. The review of size effect on stock returns covers the period 1981 to 2008.
Table 2.3: Review of major studies on size effectsStudy Major findingsBanz (1981) Small firms, on average, have significantly larger risks adjusted returns than large
firms.Fama and French (1992)
Size (ME) and book-to-market equity (BE/ME) provide a simple and powerful characterization of the cross-section of average stock returns.
Fama and French (1993)
Portfolios constructed to mimic risk factors related to size and BE/ME add substantially to the variation in stock returns explained by a market portfolio.
Daniel and Titman (1997)
There is no evidence of a separate distress factor and, it is characteristics (size & book-to-market) rather than factor loadings that determine expected returns.
Daniel, et.al (2001)
In equilibrium, there is ability of fundamental/price ratios and market value to forecast stock returns, and the domination of beta by these variables.
Vassalou and Xing (2004)
Both the size and book-to-market effects can be views as default effects which are in sum the case of size effect.
Fama and French (2008)
The anomalous returns associated with net stock issues, accruals, and momentum are pervasive; they show up in all size groups (micro, small, and big) in cross-section, and they are also strong in sorts, at least in the extremes.
The relationship between total market value of equity and common stock returns is
examined by Banz (1981). The study covered the observations from 1926 to 1975, and
included all common stocks listed in the NYSE. Data were derived from monthly returns
file of the CRSP, University of Chicago. Using pooled cross-sectional and time series
regression, the study reported that small NYSE firms, on average, have significantly
larger risks adjusted returns than large NYSE firms. The evidence suggested that the
CAPM is not correctly specified. However, the size is not linear in the market protection
but is most pronounced for the smallest firms in the sample. The effect is not very stable
through time. An analysis of the ten year sub-period showed substantial differences in the
magnitude of the coefficient of the size factor. Finally, the study concluded that there is
no theoretical foundation for such an effect, and it is not confirmed whether the factor is
size itself or whether size is just a proxy for one or more true but unknown factors
correlated with size. Therefore, the study reasoned that it is possible, however to offer
some conjectures and even discuss some factors for which size is suspected to proxy.
The observations starting from July 1963 to December 1990, Fama and French (1992)
conducted a analysis on the cross-section of expected stock returns. In area of portfolio
23
management several studies have been undertaken to specify the characteristics of stock
returns. Among the others, CAPM is the most popular model uses a single factor, beta, to
compare a portfolio with the market as a whole. Then, some research findings showed
contradictory results in the literature of finance with CAPM. The motivation of this
research is guided by such evidences. The purpose of the study is to evaluate joint roles of
market beta, size, earning yield, leverage, and Book to Market Equity in the cross section
of average stock returns on NYSE, AMEX, and NASDAQ stocks. The study capture the
cross-sectional variation in average stock returns associated with size (ME) and book-to-
market equity. Sample included are all non-financial firms listed in NYSE, AMEX and
NASDAQ and accounting information were collected from CRSP and COMPUSTAT
database. Fama and MacBeth (1973) regression is used for the analysis. The study
revealed strong relationship between the average stock returns and size, but there was no
reliable relation between average returns and beta. When the stock returns is sorted based
on earnings yield, a familiar U-shape relation is observed. The relation between average
returns and book-to-market equity is strongly positive. The FM regressions also
confirmed the importance of book-to-market equity in explaining the cross section of
average stock returns. This book to market equity relation is found stronger than the size
effects when both size and book-to-market equity were included in multivariate
regressions. The author reported book-to-market equity is consistently the most powerful
factor explaining the cross-section of average stock returns, whereas size effect was found
weaker. Based on the regression results and analysis of portfolios, the study concluded
that size and book-to-market equity provide a simple and powerful characterization of the
cross-section of average stock returns.
The common five risk factors for stocks and bonds returns are idenified by Fama and
French (1993). Three factors: an overall market factors, factor related to firm size and
book-to-market equity are the stock market related factors. For instance if the portfilos
constructed on mimic risk factors related to size and BE/ME, capture a strong common
variation in returns as the evidence that size and book-to-market equity indeed proxy for
sensitivity to common risk factors in stock returns. Other two bond market factors: default
risks, and factor related to maturity, are the bond risk factors. Stock returns have shared
variation due to the stock market factors, and which are linked to the bond returns through
shared variation in the bond market factors. Mostly the bond market factors capture the
common variation in bond returns, except for low-grade corporates. Most importantly,
these common risk factors seem to explain average returns on stocks and bonds. On the
24
other hand, variables that have no special standing in asset pricing theory shows reliable
power to explain the cross-section of average returns. The list of empirically determined
average stock returns variables includes size (ME, stock price times number of shares),
leverage, earning price ratio (E/P), and book-to-market equity (the ratio of the book value
of a firm’s common stock, BE to its market value, ME). The study employed the time-
series regression approach of Black, et.al (1972). Monthly returns on stocks and bonds are
regressed on the returns to a market portfolio of stocks and mimicking portfolios for size,
book-to-market equity, and term-structure risk factors in returns. The time-series
regression slopes are factor loadings that are unlike size or book-to-market equity have a
clear interpretation as risk-factor sensitivities for bonds as well as for stocks. Thus, Fama
and French (1993) confirm that portfolios constructed to mimic risk factors related to size
and BE/ME add substantially to the variation in stock returns explained by a market
portfolio. Moreover, a three-factor asset-pricing model that includes a market factor and
risk factors related to size and BE/ME seems to capture the cross-section of average
returns on U.S. stocks.
There is now considerable evidence that the cross-sectional pattern of stock returns can be
explained by characteristics such as size, leverage, past returns, dividend-yield, earnings-
to-price ratios, and book-to-market ratios (Fama and French, 1993). The study argued that
the association between these characteristics and returns arise because the characteristics
are proxies for non-diversifiable factor risk. Whereas, Fama and French (1992, 1996)
examine all of these variables simultaneously and concluded that with the exception of
the momentum strategy described by Jegadeesh and Titman (1993) the cross-sectional
variation in expected returns can be explained by only two of these characteristics, size
and book-to-market. Firm sizes and book-to-market ratios are both highly correlated with
the average returns of common stocks. In contrast, the evidence of the study indicates that
the return premia on small capitalization and high book-to-market stocks does not arise
because of the co-movements of these stocks with pervasive factors. It is the
characteristics rather than the covariance structure of returns that appear to explain the
cross-sectional variations in stock returns. The study focus on the factor portfolios
suggested by Fama and French (1993) and draw the conclusion that factor loadings
measured with respect to the various macro factors used by Chan, et.al (1985), Chen, et.al
(1986), and Jagannathan and Wang (1996) also failed to explain the stock returns once
characteristics are taken into account. Thus, implying different forms of regression
models, portfolios analysis and analysis of factor loadings, Daniel and Titman (1997)
25
demonstrated two major things: First, there is no evidence of a separate distress factor.
Most of the co-movement of high book-to-market stocks is not due to distressed stocks
being exposed to a unique distress factor, but rather, because stocks with similar factor
sensitivities tend to become distressed at the same time. Second evidence suggests that it
is characteristics (size & book-to-market) rather than factor loadings that determine
expected returns. It shows that factor loadings do not explain the high returns associated
with small and high book-to-market stocks beyond the extent to which they act as proxies
for these characteristics.
Daniel, et.al (2001) offered a model in which asset prices reflect both covariance risk and
misperceptions of firms’ prospects, and in which arbitrageurs’ trade against mispricing.
The classical theory of security market equilibrium is based on the interaction of fully
rational optimizing investors. Several important studies have been explored alternatives to
the premise of full rationality in recent years. One approach model market misevaluation
as a consequence of noise or positive feedback trades. Another approach analyzes how
individuals form mistaken beliefs or optimize incorrectly, and derives the resulting trades
and misevaluation. The objective of the study is to offer a theory of asset pricing in which
the cross section of expected security returns is determined by risk and investor
misevaluation. In equilibrium, expected returns are linearly related to both risk and
mispricing measures e.g., fundamental/price ratios. With many securities, mispricing of
idiosyncratic value components diminishes but systematic mispricing does not. The
theory offer untested empirical implications about volume, volatility, fundamental/price
ratios and mean returns which is consistent with several empirical findings. Thus, the
study included that the ability of fundamental/price ratios and market value to forecast
stock returns, and the domination of beta by these variables.
A firm is said to be default when it fails to service its debt obligations. Therefore, default
risk induces lenders to require from borrowers a spread over the risk-free rate of interest.
This spread is an increasing function of the probability of default of the individual firm.
Vassalou and Xing (2004) estimated the default likelihood indicators (DLI) for individual
firms using equity beta. The main purpose of the study is to address the issue that
investors are still known very little about how default risk affects equity returns. The DLI
are nonlinear functions of the default probabilities of the individual firms and are
calculated using the contingent claims methodology of Black and Scholes (1973) and
Merton (1974). The study used the COMPUSTAT file of all firms for the analysis starting
from January 1971 to December 1999. The major findings of the study are: the measure
26
of default risk contains very different information from the commonly used aggregate
default spreads which is default risk, intimately related to the size and book-to-market
characteristics of a firm. It shows that both effects are intimately related to default risk.
Small firms earn higher returns than big firms, only if they also have higher default risk.
Similarly, value stocks earn higher returns than growth stocks, if their risk of default is
high. In addition, high-default-risk firms earn higher returns than low default risk firms,
only if they are small in size and/or high book-to-market equity. In all other cases, there is
no significant difference in the returns of high and low default risk stocks. With these
findings the study concluded that both the size and book-to-market effects can be views
as default effects which is in sum the case of size effect.
In a study of dissecting anomalies, Fama and French (2008) considered the patterns of
average stock returns which do not explained by CAPM. Two approaches were used to
identify anomalies: sorts of returns on anomaly variables, and regressions, in the spirit of
Fama and MacBeth (1973) to explain the cross-section of average returns. The data
collection started from at the end of each June 1963 to end with 2005. NYSE, Amex, and
NASDAQ stocks were allocated into three size groups - microcaps (tiny), small stocks,
and big stocks. The breakpoints are the 20th and 50th percentiles of end-of-June market
cap for NYSE stocks. The findings of the study includes: as the previous work found that
net stock issues, accruals, momentum, profitability, and asset growth are associated with
anomalous average returns. Smilarly, the study explored the pervasiveness of these return
anomalies via sorts and cross-section regressions estimated separately on microcaps,
small stocks, and big stocks. The book-to-market ratio, net stock issues, accruals, and
profitability all produce average regression slopes that are indistinguishable across size
groups. The measured net of the effects of size and B/M, the equal- and value-weight
abnormal hedge portfolio returns associated with momentum, net stock issues, and
accruals are strong for all size groups (and thus pervasive). There is a more serious stain
on the net stock issues anomaly. The regression results showed that, at least for 1963 to
2005, each of the anomaly variables seems to have unique information about future
returns. All the anomaly variables are at least rough proxies for expected cash flows.
Finally, the study commonly interprets the average returns associated with anomaly
variables as evidence of market inefficiency. In sum, the anomalous returns associated
with net stock issues, accruals, and momentum are pervasive; they show up in all size
groups (micro, small, and big) in cross-section regressions, and they are also strong in
sorts, at least in the extremes. The asset growth and profitability anomalies are less
27
robust. There is an asset growth anomaly in average returns on microcaps and small
stocks, but it is absent for big stocks. Among profitable firms, higher profitability tends to
be associated with abnormally high returns, but there is little evidence that unprofitable
firms have unusually low returns.
b) Review of major studies on stock returns analysis, return decomposition and methodology effects
The financial investment focus towards the returns in terms of shareholders wealth
maximization or simply, on financial returns. The returns on investment is not an isolated
terms, it is relative and interrelated with multiple factors including its own behavior. This
section includes the major studies on stock returns which help to analyze the stock returns
in depth. The stock return analysis along with its decomposition and the methodological
effects for the period 1972 to 2008 have been presented herein first and second sub-
section respectively.
i) Review of major studies on stock returns
The level of market efficiency is formed based on the speed of adjustment of new
information. Among the others, the market information is one that causes the stock
returns. The market would be consistent if there is strong form of efficiency but the strong
form of efficiency is imaginary so that the stock returns moves ups and downs as per the
information as well as on the basis of time being. Table 2.4 shows the major studies on
stock returns analysis.
Rendleman et.al (1982) aimed to reexamine the previous study (Reinganum's study)
which indicates that abnormal returns could not be earned unexpected quarterly earnings
information, and documented precisely the response of stock prices to earnings
announcements. The study used a very large sample of stocks and daily returns which
represents the most complete and detailed analysis of quarterly earnings. The major
findings of the study is contrary to those of the earlier study and showed that abnormal
returns could have been earned almost any time during the 1970's. The analysis also
indicated that risk adjustments matter little in this type of work. Finally, the study found
roughly 50 percent of the adjustment of stock returns to unexpected quarterly earnings
occurs over a 90-day period after the earnings are announced.
Table 2.4: Review of major studies on stock returnsStudy Major findingsRendleman, et.al. (1982)
Abnormal returns could have been earned almost any time. The analysis also indicated that risk adjustments matter little in this type of work.
28
Poterba and Summers (1988)
Positive autocorrelation in returns over short horizons and negative autocorrelation over longer horizons, although random-walk price behavior cannot be rejected at conventional statistical levels. With this, the conclusion is substantial movements in required returns are needed to account for the correlation patterns.
Kothari, et.al (1995)
The relationship between book-to-market equity and returns is weaker and less consistent.
Fama and French (1996b)
Except for the continuation of short-term returns, the anomalies largely disappear.
Fama and French (1996a)
Survivor bias does not explain the relation between book-to-market equity and average returns and, beta alone cannot explain average returns.
Fama and French (1997)
The costs of equity for industries are imprecise.
Devas, et.al (2000)
The value premium in average stock returns in US is robust.
Asness et.al (2000)
Within-industry momentum has predictive power for the firm’s stock return beyond that captured by across-industry momentum.
Fama and French (2002)
The average stock return on the last half-century is a lot higher than expected in US.
Malmendier and Tate (2008)
Overconfident CEOs over-estimate their ability to generate returns.
The transitory components in stock prices are investigated by Poterba and Summers
(1988). After showing that statistical tests have little power to detect persistent deviations
between market prices and fundamental values, the study considered whether prices are
mean-reverting. The study is based on the data from the United States and 17 other
countries. The point estimates of the empirical work explain the positive autocorrelation
in returns over short horizons and negative autocorrelation over longer horizons,
although random-walk price behavior cannot be rejected at conventional statistical levels.
The authorities indicated that substantial movements in required returns are needed to
account for the correlation patterns. The study also discussed with persistent, but
transitory, disparities between prices and fundamental values.
A study by Kothari, et.al (1995) examined whether beta explains cross-sectional variance
in average returns over the post-1940 periods as well as the longer post-1926 period, and
whether book-to-market equity captures cross-sectional variations in average returns over
a longer 1947 to 1987 period. The authors noted that the relationship between book-to-
market equity and returns is weaker and less consistent than that in Fama and French
(1992). They claimed that past book-to-market results using COMPUSTAT data are
affected by a selection bias and provide indirect evidence. Using an alternative data
sources from standard poor’s industry level from 1947 to 1987, the authors have noted
that book-to-market is at best weakly related to average stocks returns. The study
presented evidence that average returns do indeed reflects sustainable compensation for
beta risk, provided that betas are measured at the annual interval. Finally, the authors
claimed that the failure of a significant relation between book-to-market equity and
29
returns to emerge the standards poor’s industry portfolios poses a serious challenge to
The study is based on previous work that average returns on common stocks are related to
firm characteristics like size, earnings/price, cash flow/price, book-to-market equity, past
sales growth, long-term past returns, and short-term past returns, Fama and French
(1996b). Because these patterns on average returns apparently are not explained by the
CAPM, they are called anomalies. The three-factor time series regression models in
Fama and French (1993), the 25 Fama and French (1993) Size-BE/ME Portfolios of
value-weighted NYSE, AMEX and NASD stocks, excess return portfolios were formed
based on Lakonishok, et.al (LSV 1994) using COMPUSTAT accounting data, LSV
double-sort portfolios, portfolios formed on past returns, one-factor CAPM excess-return
regressions and alike rigorous models and procedures were used for the analysis for the
30 years of data covering 1964 to 1993. Fama and French (1993) found that the three-
factor risk-return relation is a good model for the returns on portfolios formed on size and
book-to-market equity. The study that also explained the strong patterns in returns
observed when portfolios are formed on earnings/price, cash flow/price, and sales growth,
variables recommended by Lakonishok, et.al (1994) and others. The three-factor risk-
return relation also captures the reversal of long-term returns documented by DeBondt
and Thaler (1985). Thus, portfolios formed on E/P, C/P, sales growth, and long-term past
returns do not uncover dimensions of risk and expected return beyond those required to
explain the returns on portfolios formed on size and BE/ME. The three-factor risk-return
relation is, however, just a model. It surely does not explain expected returns on all
securities and portfolios. The study found that cannot explain the continuation of short-
term returns documented by Jegadeesh and Titman (1993) and Asness (1994). Thus, the
study concluded, except for the continuation of short-term returns, the anomalies largely
disappear in a three-factor model. The results are consistent with rational ICAPM or APT
asset pricing, but also consider irrational pricing and data problems as possible
explanations.
Fama and French (1996a) revealed that survivor bias does not explain the relation
between book-to-market equity and average returns. The study used COMPUSTAT data
from the period 1928 to 1993. The portfolios in June of each year were formed using
betas on the NYSE value-weight market portfolio estimated with two to five years of past
monthly returns. The result showed that the average monthly and annual post formations
returns initially increased with post formation betas, but relation between average returns
30
and beta was rather flat from fourth to tenth beta deciles. However, the authors have also
explained that univariate beta regressions leave an unexplained size effect. In the
portfolios formed on size and beta, the average beta premiums form univariate
regressions of return on beta underestimated the positive relation between beta and
average returns produced by size sort and overestimated the relation between beta and
average returns produced by beta sort. Therefore, result suggested that beta alone cannot
explain average returns.
The study estimated that costs of equity for industries are imprecise, Fama and French
(1997). The standard errors of more than three percent per year are typical for both the
CAPM and the three-factor model of Fama and French (1993). The study found that
these large standard errors are the result of uncertainty about true factor risk premiums
and imprecise estimates of the loadings of industries on the risk factors. Thus, the
estimates of the cost of equity for firms and projects are surely even less precise.
The study documented that the value premium in U.S. stock returns is robust (Devas,
et.al, 2000). The positive relationship between average returns and book-to-market equity
and the three-factor risk model explains the value premium better than the hypothesis that
the book-to-market characteristic is compensated irrespective of risk loadings. The study
is based on data from 1929 to 1997, derived from Moody’s industrial manuals and
COMPUSTAT. Sample firms were selected from the NYSE, AMEX and NASDAQ
industrials and non-industrials. Fama and French (1993) three-factor asset pricing model
and characteristics model are employed. The findings showed that the value premium in
average stock returns is robust. The three-factor model explains the value premium better
than the characteristics model. Finally, when portfolios are formed from independent
sorts of stocks on size and BE/ME, the three-factor model is rejected. Based on these
results, the study concluded that the three-factor model is just a model and thus an
incomplete description of expected returns.
Within-industry momentum has predictive power for the firm’s stock returns beyond that
captured by across-industry momentum and a significant short-term (one month) industry
momentum effect which remains strongly significant when restrict the sample to only the
most liquid firms (Asness, et.al 2000). The study considered the sample of all firms listed
on the NYSE, AMEX, and NASDAQ stock exchanges from July 1963 through
December 1998 and the necessary financial data were retrieved from COMPUSTAT
database. Fama-MacBeth regression model and its modified models along with two-way
31
sorts of portfolios and descriptive statistics are employed for the analysis. Originally
established by Fama and French (1997), sample firms are categorized into 48 industries.
To explore the better proxies for the information about future returns contained in firm
characteristics such as size, book-to-market equity, cash flow-to-price, percent change in
employees, and various past returns measure were obtained by breaking these
explanatory variables into two industry-related components. The first component is the
difference between firms’ own characteristics and the average characteristics of their
industries i.e. within-industry variables and, the second is average characteristics of
firms’ industries i.e. across-industry variables. In conclusion, the study provided the
better way of sorting stocks and primarily, within-industry and across-industry variables
are better able to explain the cross-section of expected stock returns than risk proxies in
the more common market-wide form.
A study is designed to estimate the equity premium using dividend and earnings growth
rates to measure the expected rate of capital gain, Fama and French (2002). The equity
premium is the difference between the expected returns on the market portfolio of
common stocks and the risk-free interest rate. Dividends and earnings are used to estimate
the expected stock returns. The explanation of the model used is: the average stock return
is the average dividend yield plus the average rate of capital gain. The CRSP value-
weighted portfolio of NYSE, AMEX and NASDAQ stocks from 1951 to 2000 are used
for the analysis. The results estimates the dividend growth rates for 1951 to 2000, 2.55
percent and earnings growth rates 4.32 percent, are much lower than the equity premium
produced by the average stock return, 7.43 percent. The evidence suggests that the high
average return for 1951 to 2000 is due to a decline in discount rates that produces a large
unexpected capital gain. Thus, the main conclusion is that the average stock returns on the
last half-century is a lot higher than expected.
A study analyzes the top level overconfidence on acquisition and its impact on market or
the market reaction. Does CEO’s overconfidence help to explain merger decisions? is the
focus of Malmendier and Tate (2008). Generally, overconfident CEOs over-estimate
their ability to generate returns. As a result, they overpay for target companies and
undertake value-destroying mergers. The effects are strongest if they have access to
internal financing. The study tests these predictions using two proxies for
overconfidence: CEOs’ personal over-investment in their company and their press
portrayal. The result shows that the odds of making an acquisition are 65percent higher if
the CEO is classified as overconfident. The effect is largest if the merger is diversifying
32
and does not require external financing. The market reaction at merger announcement is
significantly more negative than for non-overconfident CEOs. The study considered
alternative interpretations including inside information, signaling, and risk tolerance
while analyzing the relationship.
ii) Review of major studies on stock return decomposition and methodology effects
In principle, decomposition is to make a complex problem into simple. It helps to get the
thinking straight into simpler way with a logical reasoning and come out with a potential
solution for the complex issue. The decomposition approach in other words, is an attempt
to obtain relatively simple interpretation for the complex issues. With the decomposition
principle, one can identify the factors affecting stock returns. Apart from identifying the
factors contributing for stock returns, the methodology used for the study is also a major
contributor for stock return anomalies. Table 2.5 shows the major studies in stock returns
decomposition and the methodology effects as follows:
Table 2.5: Review of major studies on stock returns decomposition & methodology effects
Study Major findingsFama (1972) Return on a portfolio can be subdivided into two parts: the return from security
selection (selectivity) and the return from bearing risk (risk).
Campbell (1991) Unexpected stock returns associated with changes in expected future dividends or expected future returns.
Fama (1998) Anomalies can be due to methodology, most long-term return anomalies tend to disappear with reasonable changes in technique used for the analysis and the anomaly is stronger for small stocks.
The evaluation of the investment performance is the crucial issue in investment
management. Number of studies has been conducted on the similar topics. Among others,
Fama (1972) suggested the methods for evaluating investment performance. The previous
works are concerned with measuring performance into two dimensions, return and risk.
The study suggested somewhat finer breakdowns of the investment performance. The
goal of the performance measure itself is just to test how good the portfolio manager is at
security analysis. That is, does the portfolio manager show any ability to uncover
information about individual securities that is not already implicit in their prices? The
basic notion underlying the methods of performance evaluation is presented, and the
returns on managed portfolios can be judged relative to those of "naively selected"
portfolios with similar levels of risk. Both the measure of risk and the definition of a
naively selected portfolio were obtained from modern capital market theory. The
conclusions of the study are: the stock returns on a portfolio can be subdivided into two
33
parts: the return from security selection (selectivity) and the return from bearing risk
(risk). The return from selectivity is defined as the difference between the return on the
managed portfolio and the return on a naively selected portfolio with the same level of
market risk.
What moves the stock returns? To get the ideas on this voluminous research question and
the heated debate issue, Campbell (1991) conducted a study on variance decomposition
for stock returns. The study present a simple way to break stock market movements into
two components; one which is associated with changes in rational expectations of future
returns is "news about future returns", and one which is not is called the "news about
future dividends". The approaches and tools used for the analysis are; arbitrary correlation
approach between the two components which is important in practice, regression analysis
to describe the evolution through time of the forecasting variables, vector autoregressive
(VAR) system used to calculate the impact that an innovation in the expected return will
have on the stock price, contemporaneous regression approach regresses stock returns on
contemporaneous innovations to variables which might plausibly affect the stock market,
univariate time-series approach studies the autocorrelation function of stock returns. The
study shows that unexpected stock returns must be associated with changes in expected
future dividends or expected future returns. A vector autoregressive method is used to
breakdown the unexpected stock returns into two components. In U.S. monthly data of
NYSE retrieved from CRSP from 1927 to 1988, one-third of the variance of unexpected
returns is attributed to the variance of changing expected dividends, one-third to the
variance of changing expected returns, and one-third to the covariance of the two
components. Changing expected returns have a large effect on stock prices because they
are persistent: a 1 percent innovation in the expected return is associated with a 4 or 5
percent capital loss. Changes in expected returns are negatively correlated with changes
in expected dividends, increasing the stock market reaction to dividend news. In the
period 1952-88, changing expected returns account for a larger fraction of stock return
variation than they do in the period 1927-51.
Consistent with the market efficiency hypothesis that the anomalies are chance results,
apparent overreaction to information is about as common as underreaction, and post-
event continuation of pre-event abnormal returns is about as frequent as post-event
reversal (Fama, 1998). Most important, consistent with the market efficiency prediction
that apparent anomalies can be due to methodology, most long-term return anomalies
tend to disappear with reasonable changes in technique used for the analysis. The three-
34
factor model of Fama and French (1993) is employed to estimate the portfolios abnormal
returns, it showed that the three-factor model is not a perfect story for average returns
and considered as the bad-model. The bad-model problem can produce spurious
anomalies in event studies. All methods for estimating abnormal returns are subject to
bad-model problems, and no method is likely to minimize bad-model problems for all
classes of events. The study provides the important general message from the initial
public offerings and seasoned equity offerings results is one caution: two approaches that
seem closely related i.e. both attempt to control for variation in average returns related to
size and BE/ME, can produce much different estimates of long-term abnormal returns.
The anomalies are largely limited to small stocks because small stocks always pose
problems in tests of asset pricing models, so that they are prime candidates for bad-model
problems in tests of market efficiency on long-term returns. Thus, the anomaly is
stronger for small stocks.
c) Review of major studies on investor behavior
The heated issue in financial literature is the behavioral effects on stock returns. The
financial literature explained that there are numerous qualitative factors that contribute for
stock market movements. The quantitative factors that can be measured but their
significance is questionable because of historic nature. The behavioral factors on the other
hand, significantly influence the stock movements. At the same time, it is very difficult to
articulate the level of its influences. The major studies on investor behavior have been
presented in this section. The study period range from 1994 to 2011.
i) Review of major studies on investor behavior before 2000
This sub-section focuses on the review of value strategies versus glamour strategies with
investor behavior, information processing, news and events responses, etc. Table 2.6
presents the review of major studies on investor behavior before 2000 as follows:
Lakonishok, et.al (1994) conducted a study on the most debatable, value strategies,
glamour strategies, investors’ extrapolation and risk which have attracted academic
attention as well. The value strategies call for buying stocks that have low prices relative
to earnings, dividends, book assets, or other measures of fundamental value. For many
years, scholars and investment professionals have argued that value strategies outperform
the market. While there are some agreements that value strategies produce higher returns,
but the interpretation of why they do so is more controversial. The objective of the study
is to shed further light on the two potential dimensions for why value strategies work.
35
Table 2.6: Review of major studies on investor behavior before 2000
Study Major findingsLakonishok, et.al (1994)
Value strategies yield higher returns than glamour strategies because these strategies exploit the suboptimal behavior of the typical investor and not because these strategies are fundamentally riskier.
Ikenberry et al. (1995)
The market responds mistakenly in initial phase of information and appeared to ignore much of the information conveyed through repurchase announcement.
Barberis, et.al (1998)
In a variety of markets, sophisticated investors can earn superior returns by taking advantage of under-reaction and overreaction without bearing extra risk.
Klibanoff, et.al (1998)
News events lead some investors to react more quickly.
Odean (1999)
The trading volume of a particular class of investors, those with discount brokerage accounts, is excessive. These investors trade excessively in the sense that their returns are, on average, reduced through trading.
First, the study examines more closely the predictions of the contrarian model. Second,
value strategies that bet against those investors who extrapolate past performance too far
into the future produce superior returns. Variables employed for the study are: past
performance is measured using information on past growth in sales, earnings, and cash
flow, and expected performance is measured by multiples of price to current earnings,
and cash flows. The sample period covered from the end of April 1963 to the end of April
1990. The sources of data are CRSP and COMPUSTAT, of NYSE and AMEX firms. The
results could potentially suffer from the Look-ahead or survivorship bias (Banz and
Breen, 1986) and Kothari, et.al, 1992) but methodology used is different from those in
other recent studies in ways that should mitigate this bias by First, do not use those
returns to evaluate strategies which appear such bias. Second, study only NYSE and
AMEX firms. Finally, report results for the largest 50 percent of firms on the NYSE and
AMEX. The selection bias is less serious among these larger firms (La Porta, 1993).
Couple of simple statistical tools; average, percentage, standard deviation along with
rigorous portfolio analysis and FM regression models is used for the analysis. The study
provides that value strategies (high B/M) yield higher returns because these strategies
exploit the suboptimal behavior of the typical investor and not because these strategies
are fundamentally riskier.
A total of 1239 open market share repurchases announced between January 1980 and
December 1990 by firms whose shares traded on the NYSE, ASE, or NASDAQ is
considered as the sample of the study, (Ikenberry et.al, 1995). For the performance
measurement, the study used the CAR approach and the buy-and-hold approach. The
long-run firm performance following open market share repurchase announcement
indicated that the average abnormal four year buy-and-hold return measured after the
36
initial announcement is 12.1 percent where as the average market response to the
announcement of an open market share repurchase is 3.5 percent. For value stocks,
companies more likely to be repurchasing shares because of undervaluation, the average
abnormal return is 45.3 percent. For repurchases announced by glamour stocks, where
undervaluation is less likely to be an important motive, no positive drift in abnormal
return is observed. Thus, at least with respect to value stocks, the market errs in its initial
response and appears to ignore much of the information conveyed through repurchase
announcement.
The motivation of the study is the recent empirical researches in Finance which have been
uncovered two families of pervasive regularities: underreaction of stock prices to news
such as earnings announcements, and overreaction of stock prices to a series of good or
bad news. For example, the underreaction evidence shows that over horizons of perhaps
one to twelve months security prices underreact to news. In an effort to fill this gap,
Barberis, et.al (1998) propose a model of investor sentiment. As a consequence, news is
incorporated only slowly into prices, which tend to exhibit positive autocorrelations over
these horizons. A related way to make this point is to say that current good news has
power in predicting positive returns in the future. The overreaction evidence shows that
over longer horizons of perhaps three to five years, security prices overreact to consistent
patterns of news pointing in the same direction. That is, securities that have had a long
record of good news tend to become overpriced and have low average returns afterwards.
Put differently, securities with strings of good performance, however measured, receive
extremely high valuations. This effort presents a parsimonious model of investor
sentiment, or of how investors form beliefs, which is consistent with the empirical
findings. The model is based on psychological evidence and produces both under-reaction
and overreaction for a wide range of parameter values. The existence of this model
challenge to the efficient markets theory because it suggests that in a variety of markets,
sophisticated investors can earn superior returns by taking advantage of under-reaction
and overreaction without bearing extra risk.
In an effort to investigate the investors’ reactions to salient news, Klibanoff, et.al (1998)
conducted a study on ‘investor reaction to salient news in closed-end country funds.’
Panel data on prices and net asset values are used to test whether dramatic country-
specific news affects the response of closed-end country fund prices to asset value.
Authors believe that anomalous empirical regularities in financial returns can be
explained by investor underreaction or overreaction to news. The study is focused on a
37
particular form of cognitive error as a source of investor over- and underreaction, and
examined the hypothesis that individual investor assign more importance to more
prominent news and assign less importance to less prominent news, even if the two pieces
of news have the same effect on fundamental value. The objective of the study is to use
financial market data to detect the salience effect on investor reaction in a non-laboratory
setting. Major news events on the front page of The New York Times (NYT) were
collected and correlate with the degree of reaction in financial asset prices. Sample of
country funds consists of the 39 single-country publicly traded funds from January 1986
through March 1994 and that have at least twelve months of price and net asset value
data, and weekly data on funds represent 25 countries. Ordinary least square regressions
were used for the analysis. The results showed that in a typical week, prices underreact to
changes in fundamentals; the (short-run) elasticity of price with respect to asset value is
significantly less than one. In weeks with news appearing on the front page of NYT,
prices react much more; the elasticity of price with respect to asset value is closer to one.
Thus, the findings of the study are consistent with the hypothesis that news events lead
some investors to react more quickly.
Based on the issue - trading volume on the financial market seems high, perhaps higher
than can be explained by models of rational markets. The study demonstrates that the
trading volume of a particular class of investors, those with discount brokerage accounts,
is excessive, Odean (1999). These investors trade excessively in the sense that their
returns are, on average, reduced through trading. Thus, the study tests the hypothesis that
investors trade excessively because they are overconfident. Overconfident investors may
trade even when their expected gains through trading are not enough to offset trading
costs. In fact, even when trading costs are ignored, these investors actually lower their
returns through trading. The study examines return patterns before and after the purchases
and sales made by these investors. The investors tend to buy securities that have risen or
fallen more over the previous six months than the securities they sell. They sell securities
that have, on average, risen rapidly in recent weeks. The study suggest that these patterns
can be explained by the difficulty of evaluating the large number of securities available
for investors to buy, by investors’ tendency to let their attention be directed by outside
sources such as the financial media, by the disposition effect, and by investors' reluctance
to sell short.
ii) Review of major studies on investor behavior during 2000s
38
After 2000, the major behavioral studies on investor behavior have been presented in
Table 2.7 in this sub-section. The studies includes investors response to public and private
signals, buying and selling behavior, behavior of men and women investors, etc as below:
There is a large amount of evidences that stock prices are predictable and that stock
returns exhibit reversal at weekly and three-to-five year intervals and drift over 12 month
periods. Chan (2003) aimed to deepen the understanding of how information flows drive
anomalies. Using a comprehensive database of headlines about individual companies, the
study examines monthly returns after two sources of stimuli. The first is public news,
which is identifiable from headlines and extreme concurrent monthly returns. The second
is large price movements unaccompanied by any identifiable news. The study, then,
examines monthly returns following public news and compares them to stocks with
similar returns but no identifiable public news. The study presents that there is a
difference between the two sets. Thus, the study concluded that there is strong drift after
bad news and investors seem to react slowly to bad news. The results is also depicted that
stock returns reversal after extreme price movements unaccompanied by public news.
Table 2.7: Review of major studies on investor behavior during 2000sStudy Major findingsChan (2003) Investors are appeared to underreact to public signals and overreact to perceived
private signals. For instance, Investors tend to react slowly to the bad news information.
Biais et.al. (2005)
Miscalibration reduces and self-monitoring enhances trading performance. The effect of the psychological variables is strong for men but non-existent for women.
Barber and Odean (2008)
Individual investors display attention-driven buying behavior, they are net buyers on high-volume days, following both extremely negative and extremely positive one-day returns, and when stocks are in the news. On the other hand, the institutional investors - especially the value-strategy investors - do not display attention-driven buying.
Kaniel, et.al (2008)
The trading of individual investors provides two important results. First, net individual trading is positively related to future short-horizon returns: Prices go up in the month after intense buying by individuals and go down after intense selling by individuals. Second, the predictive ability of net individual trading with respect to returns is not subsumed by volume or the return reversal phenomenon.
Foucault, et.al (2011)
The effect of retail trading on volatility is positive, the positive effect is consistent with the view that some retail investors behave as noise traders.
Sun and Wei (2011)
Analysts make more judgment-intensive decisions, such as issuing stock recommendations; they overweight intangible information, leading to overreaction to intangible information. On the contrary, when analysts make less judgment-intensive decisions, such as earnings per share (EPS) forecasts, there is no such evidence of overreaction. The study supports the hypothesis that investors are overly sensitive to intangible information when they need to make more subjective judgments.
Doskeland and Hvide (2011)
Individuals with a comparative advantage in collecting information can obtain asymmetric information and earn abnormal returns. On the other hands, investors have a preference for professionally close stocks even if such holdings generate negative abnormal returns.
39
The degree of overconfidence in judgment in the form of miscalibration is measured by
Biais, et.al. (2005). The tendency to overestimate the precision of one's information and
self-monitoring, a form of attentiveness to social cues of 245 participants and also
observe their behavior in an experimental financial market under asymmetric information.
Miscalibrated traders, underestimating the conditional uncertainty about the asset value,
are expected to be especially vulnerable to the winner's curse. High self-monitors are
expected to behave strategically and achieve superior results. Thus, the study concluded
that miscalibration reduces and self-monitoring enhances the trading performance. The
effect of the psychological variables is strong for men but non-existent for women.
The hypothesis that individual investors are net buyers of attention grabbing stocks, e.g.,
stocks in the news, stocks experiencing high abnormal trading volume, and stocks with
extreme one-day returns is evaluated and confirmed by Barber and Odean (2008).
Attention-driven buying results from the difficulty that investors have searching the
thousands of stocks they can potentially buy. Individual investors do not face the same
search problem when selling because they tend to sell only stocks they already own. The
study hypothesize that many investors consider purchasing only stocks that have first
caught their attention. Thus, preferences determine choices after attention has determined
the choice set. The study collected the required the data from four sources: a large
discount brokerage - records for the investments of 78,000 households from January 1991
through December 1996, a small discount brokerage - daily trading records from January
1996 through 15 June 1999 and 14,667 accounts for individual investors, a large full-
service brokerage - investments of households for the 30 months ending in June 1999,
and the Plexus Group—a consulting firm that tracks the trading of professional money
managers for institutional clients, provide the daily trading records for 43 institutional
money managers and span the period January 1993 through March 1996. During the
sample period, only 7194297 common stock trades are included for the analysis out of 10
million total trades: 3,974,998 purchases with a mean value of $15,209 and 3,219,299
sales with a mean value of $21,169. In sum, consistent with the predictions, the study
concluded that individual investors display attention-driven buying behavior, they are net
buyers on high-volume days, following both extremely negative and extremely positive
one-day returns, and when stocks are in the news. Attention-driven buying is similar for
large capitalization stocks and for small stocks. On the other hand, the institutional
investors - especially the value-strategy investors - do not display attention-driven buying.
40
For a variety of reasons, financial economists tend to view individuals and institutions
differently. In particular, while institutions are viewed as informed investors, individuals
are believed to have psychological biases and are often thought of as the proverbial noise
traders. Kaniel, et.al (2008) investigates the dynamic relation between net individual
investor trading and short-horizon returns for a large cross-section of NYSE stocks. The
sample contained all common domestic stocks that were traded on the NYSE any time
between January 1, 2000, and December 31, 2003. The analysis of the trading of
individual investors on the NYSE provides two important results. First, net individual
trading is positively related to future short-horizon returns: Prices go up in the month after
intense buying by individuals and go down after intense selling by individuals. Second,
the predictive ability of net individual trading with respect to returns is not subsumed by
volume or the return reversal phenomenon.
Foucault, et.al (2011) studied on retail trading activities that have a positive effect on the
volatility of stock returns, which suggests that retail investors behave as noise traders.
Anything that changes the amount or character of noise trading will change the volatility
of price (Black, 1986). The study focuses on the issue: whether retail trading has a
positive effect on volatility which is yet to be answered. The study used the database
which provides the daily returns and daily trading volumes for each stock listed on the
French stock market from September 1998 to September 2002. The sample for the study
is 678 stocks in the control group with standard deviation 55 and 155 stocks in the treated
group with standard deviation 5 in each month. The study analyzed the reform of the
French stock market to assess the effect of retail investors on volatility. The reform makes
trading relatively more costly for retail investors in a subset of listed stocks and triggers a
drop in retail trading for these stocks relative to stocks unaffected by the reform. The
study found that the volatility of the stocks affected by the reform declines after the
implementation of the reform, relative to other stocks, which means that the effect of
retail trading on volatility is positive. The argument is that the positive effect is consistent
with the view that some retail investors behave as noise traders. In support of this claim,
the evidence showed that the reform also triggers a drop in the size of price reversals and
the price impact of trades for the stocks affected by the reform.
In order to relate the intangible information and analyst behavior, Sun and Wei (2011)
documented the direct evidence that when analysts make more judgment-intensive
decisions, such as issuing stock recommendations, they overweight intangible
information, leading to overreaction to intangible information. On the contrary, when
41
analysts make less judgment-intensive decisions, such as earnings per share (EPS)
forecasts, there is no such evidence of overreaction. More specifically, analyst
recommendations are much more sensitive to intangible information, while EPS forecasts
are more sensitive to tangible information. The sensitivity of long-term growth forecasts
to intangible and tangible information fall in between. The study also test and found that
the overconfidence bias in analyst recommendations contributes to the market
overreaction to intangible information. The results are consistent with the overconfidence
hypothesis which suggests that investors should be overly sensitive to intangible
information when they need to make more subjective judgments.
As time goes on, there are more and more convincing evidence that the right method in
investments is to put fairly large sums into enterprises which one thinks one knows
something about and in management of which one thoroughly believes (J.M. Keynes).
Based on this statement, Doskeland and Hvide (2011) hypothesize that professional
proximity is a route through which individuals can have a comparative advantage in
collecting information. The proximity can give a false feeling of competence on the one
hand or the access to value-relevant information that can lead to abnormally high returns
on the other. Popular belief suggests that some individuals have asymmetric information
and can gain from being undiversified (Merton (1987)). For instance, Warren Buffet who
has been generating strong investment performance using this approach over a 30-year
period (e.g., Martin and Puthenpurackal, 2008) advises “Invest within your circle of
competence. It’s not how big the circle is that counts, it’s how well you define the
parameters” (Fortune, November 11, 1993). The study employed the common stock
transactions of all Norwegian individual investors at the Oslo Stock Exchange (OSE) over
a 10-year period. The data set combined the full trade records of each individual with
exceptionally detailed socio-demographic information at a yearly level over a 20-year
period i.e. yearly panel of work history for each individual, including the industry and
ticker code of their employer, where investors live, geographically proximate, etc.
Further, the study defined the “expertise” stocks which SIC code matches the two-digit
SIC code of the individual’s employer. The findings of the study suggest that individuals
overweigh their holdings in expertise stocks. Individuals spend much of their time
building and maintaining their professional career, and thus they gain a considerable
amount of industry-specific experience. Accordingly, the study conjectured that
professional proximity is a route through which individuals can obtain a comparative
advantage in acquiring value-relevant information and hence realize abnormal stock
42
market returns. Professionally close investments is a particularly fitting environment to
detect abnormal returns following conventional portfolio theory, since investors should
invest in professionally close investments only if they are informed. These findings
provide clear evidence of a behavioral bias in individuals’ investment choices.
Overconfidence seems to be the most likely explanation for why individuals trade in
professionally close stocks. The lack of any evidence of abnormal returns for a very
plausible candidate suggests that individual investors are not able to profit from
asymmetric information. Another take-away of the results is to provide guidance to
individual investors themselves. Conventional portfolio theory recommends that investors
shy away from professionally close stocks unless they have superior information, since
such stocks carry extra risk. The study also found that investors have a preference for
professionally close stocks even if such holdings generate negative abnormal returns. It
thus seems that individual investors themselves are not aware of their poor investment
choices.
d) Review of major studies on initial public offerings
The study on initial public offerings shows the different reasons that the firm do not want
to issue IPOs when it is viewed from firms’ point of view. Whereas, taking the individual
perspectives, some studies focuses that investing in IPOs is riskier than investing in
secondary market. Further some studies position that the maintenance of target debt-
equity ratio while issuing IPOs as well as while making the repurchase decisions. Table
2.8 shows the major findings of some prominent studies on IPO issues. The study period
covers 1995 to 2006 as follows:
Table 2.8: Review of major studies on IPOs Study Major findingsLoughran and Ritter (1995)
Investing in firms issuing stock is hazardous to wealth.
Armen et.al. (2001)
The deviation between the actual and the target debt ratios plays a more important role in the repurchase decision than in the issuance decision and, when firms adjust their capital structure; they tend to move toward a target debt ratio.
Brau and Fawcett (2006)
The main reason for remaining private or do not issuing IPO is to preserve decision-making control and ownership.
The study on the new issues puzzle, Loughran and Ritter (1995) covered the sample
period from 1970 to 1990. The sample of 4753 companies going public in US stock
exchanges is analyzed. The required data were collected from CRSP and listed in Nasdaq
or Amex and NYSE daily tapes. It is shown that companies issuing stocks whether an
43
initial public offering or a seasoned equity offering, significantly underperform relative to
non-issuing firms for five years after the offering data. During the five years after the
issue, investors have received average returns of only 5 percent per year for companies
going public and only 7 percent per year for companies conducting a seasoned equity
offer. The study followed the same pattern as previous studies which evidence that firms
going public subsequently underperform and the same pattern holds for firms conducting
SEOs is new. The magnitude of the underperformance is economically important: based
upon the realized returns, an investor would have had to invest 44 percent more money in
the issuers than in non-issuers of the same size to have the same wealth five years after
the offering date. Surprisingly, this number is the same for both IPOs and SEOs. The
study also found only a modest portion of the underperformance of issuing firms can be
explained as a manifestation of book-to-market effects. Another finding is that extreme
winners that do not issue equity dramatically outperform extreme winners that do issue.
Also documented that the degree to which issuing firms underperform varies over time:
firms issuing during years when there is little issuing activity do not underperform much
at all, whereas firms selling stock during high-volume periods severely underperform. For
the analysis of data, three different procedures were used. First, calculates t-statistics
using annual holding-period returns on issuing firms relative to non-issuing firms.
Second, calculates t-statistics using a time series of cross-sectional regressions on
monthly individual firm returns and the third procedure was to calculate t-statistics using
3-factor time-series regressions of monthly returns for portfolios of issuing and non-
issuing firms. All three procedures result in rejection of the null hypothesis of no
underperformance at high degrees of statistical significance. Thus, the study concluded
that investing in firms issuing stock is hazardous to wealth.
Armen, et.al. (2001) analyzes the debt-equity choice and documented that when firms
adjust their capital structures, they tend to move toward a target debt ratio that is
consistent with theories based on tradeoffs between the costs and benefits of debt. In
contrast to previous empirical works, the study explicitly account for the fact that firms
may face impediments to movements toward their target ratio, and that the target ratio
may change over time as the firm's profitability and stock price changes. The study
conducted a separate analysis of the size of the issue and repurchase transactions suggests
that the deviation between the actual and the target ratio plays a more important role in
the repurchase decision than in the issuance decision.
44
A survey on Initial Public Offerings, Brau and Fawcett (2006) analyze the responses of
336 chief financial officers (CFOs) out of 1266 valid contact information of the selected
1500 nonfinancial private firms based on 2002 revenues, to compare practice to theory in
the areas of IPO motivation, timing, underwriter selection, under-pricing, signaling, and
the decision to remain private. The study found that the primary motivation for going
public is to facilitate acquisitions. Further, CFOs base IPO timing on overall market
conditions are well informed regarding expected under-pricing, and feel under-pricing
compensates investors for taking risk. The sample was selected based on information
from January to December 2002 and the mailed survey was conducted on May 5, 2003,
June 11, 2003 and September 12, 2003. The overall response rate for the study is 18.8
percent. The most important positive signal is past historical earnings, followed by
underwriter certification. CFOs have divergent opinions about the IPO process depending
on firm-specific characteristics. Finally, the study documented that the main reason for
remaining private is to preserve decision-making control and ownership.
e) Review of major studies on market behavior
The review of literature which is concern towards the market behavior comprises: market
predictability, investing in a certain day in a week or the weekly cycle, the seasonal
patterns, market reversal, etc have been organized in this section. The review of major
studies starts from 1965 and end with 2002. These previous studies are broadly classified
into two sub-sections based on prior 1990 and 1990 onwards.
i) Review of major studies on market behavior prior 1990
Table 2.9 presents some major contribution on market behavior prior 1990. Many people
in the investment community still believe on the market cycle, some others empirical
evidences indicates that the existence of seasonal patterns as well as weekly and daily
which are considered an essential prerequisite for investment performance. The major
findings of the studies on market behavior prior 1990 have been presented as follows.
Table 2.9: Review of major studies on market behavior prior 1990
Study Major findings
Fama (1965) The behavior of stock prices is not predictable as it follow the random walk. The chart reading through perhaps an interesting pastime but there is no real value to stock market investors.
French (1980) The study proposed the negative investing strategy - buying Monday and selling them in Friday which generates the profit even in negative daily returns of Monday, most importantly inconsistent with the general perception of positive daily returns for Monday the study documented that Monday stock returns is negative due to the weekend effect.
45
Brown and Warner (1985)
Using the simulation procedures with actual daily data, the major conclusion drawn from the study is: tests which assume non-zero cross-sectional dependence are only about half as powerful and usually no better specified than those employed assuming independence.
Ritter (1988) The ratio of stock purchases to sales by individual investors displays a seasonal pattern, with individuals having a below-normal buy/sell ratio in late December and an above-normal ratio in early January.
Schwert (1989) The average level of volatility much higher during recession, weak evidence that macroeconomic volatility helps to predict stock and bond return volatility, somewhat stronger evidence that financial asset volatility helps to predict future macroeconomic volatility, financial leverage affects stock volatility, but the effect is small, and the positive relation between trading activity and stock volatility.
For many years, the following question has been a source of continuing controversy in
both academic and business circle: to what extent can the past history of a common
stock's price be used to make meaningful predictions concerning the future price of the
stock? Fama (1965) analyzed whether the history repeats itself in that patterns of past
price behavior will tend to recur in the future or follow the random walk – the future path
of the price level of a security is no more predictable than the path of a series of
cumulated random numbers. The purpose of the study is to discuss first in more detail the
theory underlying the random-walk model and then to test the model's empirical validity.
The data consist of daily prices for each of the thirty stocks of the Dow-Jones Industrial
Average (DJIA). The time periods vary from stock to stock but usually run from about the
end of 1957 to September 26, 1962. The final date is the same for all stocks, but the initial
date varies from January, 1956 to April, 1958 so that there are thirty samples with about
1,200-1,700 observations per sample. Using frequency distribution, normal curve graphs,
portfolio formation and variance comparison, the study concluded that the behavior of
stock prices is not predictable as it follow the random walk. The chart reading through
perhaps an interesting pastime but there is no real value to stock market investors.
The process generating stock returns has been one of the most popular topics of research
in finance since Bachelier’s pioneering article, published in 1900. French (1980)
examines the process of generating stock returns by comparing the returns of different
days of the week. Ignoring holidays, the returns reported for Monday represent a three-
calendar-day investment, from the close of trading Friday to the close of trading Monday,
while the returns for other days reflect a one-day investment Therefore, if the expected
return is a linear function of the period of investment, measured in calendar time, the
mean return for Monday will be three times the mean for the other days of the week. The
findings of tests using the daily returns to the Standard and Poor’s (S&P) composite
portfolio, consisting 500 the largest firms and the total observations 6024 on NYSE from
46
1953 to 1977, are surprising. Inconsistent with the calendar and trading time models, the
mean return for Monday is significantly negative in each of five five-year sub-periods, as
well as over the full period. Thus, the study proposed the negative investing strategy -
buying Monday and selling them in Friday which generates the profit even in negative
daily returns of Monday, most importantly inconsistent with the general perception of
positive daily returns for Monday, the study documented that Monday’s stock returns is
negative due to the weekend effects.
The properties of daily stock returns and how the particular characteristics of these data
affect event study methodologies for assessing the share price impact on firm-specific
events is examined by Brown and Warner (1985). The study characterized number of
potentially important problems while using daily data like: non-normality, non-
synchronous trading and market model parameter estimation, variance estimation issues,
etc. Two hundred and fifty samples of 50 securities are selected for the analysis, the
securities were selected at random and with replacement, and the daily return data were
abstracted from CRSP files. The sample period covered on each trading days from July 2,
1962 through December 31, 1979. Using the simulation procedures with actual daily data,
the major conclusion drawn from the study is: tests which assume non-zero cross-
sectional dependence are only about half as powerful and usually no better specified than
those employed assuming independence.
In recent years, numbers of anomalies have been discovered in stock returns, among them
the turn-of-the-year effect have been generating the greatest interest. The average returns
on low-capitalization stocks are unusually high relative to those on large-capitalization
stocks in early January, a phenomenon is known as the turn-of-the-year effect. Ritter
(1988) found that the ratio of stock purchases to sales by individual investors displays a
seasonal pattern, with individuals having a below-normal buy/sell ratio in late December
and an above-normal ratio in early January. Year-to-year variation in the early January
buy/sell ratio explains forty-six percent of the year-to-year variation in the turn-of-the-
year effect during 1971-1985. The types of data employed for the study are: daily buy/sell
ratios of the cash account customers of the Merrill Lynch, Pierce, Fenner and Smith and
use the ratio of purchases and sales as a measure of the net buying activity of individual
investors. The ratios, five-scale descriptive statistics, mean difference in daily returns,
portfolio, histogram and the Ordinary Least Square (OLS) regression models are used for
the analysis. The study concluded that December’s net selling abruptly switches to net
buying at the turn of the year as well as explained why small stocks do well at the turn of
47
the year which validate the earlier finding that small stock outperform the large stock at
the end of the year.
The study analyzed the relation of stock volatility with real and nominal macroeconomic
volatility, economic activities, financial leverage, and stock trading activity, Schwert
(1989). The study employed the monthly data from 1857 to 1987. The estimates of the
standard deviation of monthly stock returns vary from two to twenty percent per month,
test for whether differences this large could be attributable to estimation error strongly
reject the hypothesis of constant variance. Regression analysis with autoregressive
moving average (ARMA), autoregressive conditional heteroscedasticity (ARCH),
distributed lag model and vector autoregressive model (VAR) are employed for the
analysis. The major conclusions of the study are: many economic series were more
volatile during depression (1929-39) particularly financial asset and industrial production;
the average level of volatility is much higher during recession; there is weak evidence that
macroeconomic volatility helps to predict stock and bond return volatility; there is
somewhat stronger evidence that financial asset volatility helps to predict future
macroeconomic volatility; financial leverage affects stock volatility but the effect is
small; and finally, the positive relationship between trading activity (both trading days
and volume) and stock volatility.
ii) Review of major studies on market behavior 1990 onwards
The relationship of market returns with factors like news, price bubbles and expectations,
the impact of research and development expenditure, market timing, etc are covered in
Table 2.10 as follows:
The rational speculation, usually presume that it dampens fluctuations caused by noise
traders. De Long, et al (1990) tried to explore on what effect do rational speculators have
on assets price? Speculators who destabilize asset prices do so by, on average, buying
when prices are high and selling when prices are low; such destabilizing speculators are
quickly eliminated from the market. By contrast, speculators who earn positive profits do
so by trading against the less rational investors who move prices away from
fundamentals. Such speculators rationally counter the deviations of prices from
fundamentals and so stabilize them. This is not necessarily the case if noise traders follow
positive feedback strategies - buy when prices rise and sell when prices fall, such
behavior common in financial markets. Theoretical arguments for the efficiency of
financial markets rely crucially on the stabilizing powers of rational speculation. Rational
48
speculators "buck the trend" and by doing so bring prices closer to fundamental values.
But, less rational investors, in another way, may pay to jump on the bandwagon and
purchase ahead of noise demand. Although, the key point to understand is that part of the
price rise is rational, part of it results from rational speculators' anticipatory trades and
from positive feedback traders' reaction to such trades. If rational speculators' early
buying triggers positive-feedback trading, then an increase in the number of forward
looking speculators can increase volatility about fundamentals. The findings of the study
generate a positive correlation of stock returns at short horizons, as positive feedback
traders respond to past price increases by flowing into the market, and negative
correlations of stock returns at long horizons, as prices eventually return to fundamentals.
Also, the study predicts that the stock market overreacts to news because such news
triggers positive feedback trading. In sum, the conclusion of the study is consistent with a
number of empirical observations about the correlation of asset returns, the overreaction
of prices to news, price bubbles, and expectations.
Table 2.10: Review of major studies on market behavior 1990 onwards
Study Major findings
De Long, et al (1990)
There is correlation of asset returns with the overreaction of prices to news, price bubbles, and expectations.
Hasbrouck (1991)
A trade's full price impact arrives only with a protracted lag; the impact is a positive and concave function of the trade size; large trades cause the spread to widen; trades occurring in the face of wide spreads have larger price impacts; and, information asymmetries are more significant for smaller firms.
Jegadeesh and Titman (1993)
The strategies which buy stocks that have performed well in the past and sell stocks that have performed poorly in the past generate significant positive returns over 3 to 12 month holding periods.
Chan, et.al (2001)
The stock price incorporates investors’ unbiased beliefs about the value of R&D expenditure.
Hirshleifer (2001)
The purely rational approach is being subsumed by a broader approach based upon the psychology of investors. In this approach, security expected returns are determined by both risk and misevaluation.
Baker and Wurgler (2002)
Market timing is an important aspect of real financing decisions.
Central to the analysis of market microstructure is the notion that in a market with
asymmetrically informed agents, trades convey information and therefore cause a
persistent impact on the security price. The magnitude of the price effect for a given trade
size is generally held to be a positive function of the proportion of potentially informed
traders in the population. Hasbrouck (1991) suggests that the interactions of security
trades and quote revisions be modeled as a vector autoregressive system. Within this
framework, a trade's information effect may be meaningfully measured as the ultimate
49
price impact of the trade innovation. The sample contained 20 firms, in each quartile
which had at least 500 transactions and the data abstracted from Institute for the Study of
Security Markets (ISSM) over the 62 trading days in the first quarter of 1989. Estimates
for a sample of NYSE issues suggest: a trade's full price impact arrives only with a
protracted lag; the impact is a positive and concave function of the trade size; large trades
cause the spread to widen; trades occurring in the face of wide spreads have larger price
impacts; and, information asymmetries are more significant for smaller firms.
A Popular view held by many journalists, psychologists, and economists is that
individuals tend to overreact to information. Jegadeesh and Titman (1993) documented
that strategies which buy stocks which performed well in the past and sell stocks which
performed poorly in the past generate significant positive returns over 3 to 12 month
holding periods. It is also found that the probability of these strategies is not due to their
systematic risk or to delayed stock price reactions to common factors. However, part of
the abnormal returns generated in the first year after portfolio formation dissipates in the
following two years. A similar pattern of returns around the earnings announcements of
past winners and losers is also documented. The study analyzed the NYSE and AMEX
stocks from 1965 to 1989 abstracted from the CRSP daily returns database. Relative
strength portfolio, one-factor model and the multi-factor models, covariance analysis,
lead-lag effects, events analysis as well as the regression models were used for the
analysis.
Whether stock prices fully value firms’ intangible assets, specially research and
development (R&D) expenditures is examined by Chan, et.al (2001). The market value of
a firm’s shares ultimately reflects the value of all its new assets. When, most of the assets
are physical, the link between asset values and stock prices is relatively apparent. But, in
modern economics, a large part of a firm’s value may reflect its intangible assets and
these are not reported in firms’ financial statements and treated as a current expenditure.
When a firm has large amounts of such intangibles, the lack of accounting information
generally complicates the task of equity valuation is the issue of the study. In particular,
R&D activity, one type of intangible asset has been the subject of much attention, is the
focus of the study. The evidence of the study does not support a direct link between R&D
spending and future stock returns. Thus, it does not appear that the average historical
stock returns of firms doing R&D outperform the returns of firms without R&D which is
consistent with the hypothesis that the stock price incorporates investors’ unbiased beliefs
about the value of R&D. However, the market is apparently too pessimistic about beaten-
50
down R&D intensive technology stocks’ prospects. The study found that companies with
high R&D to equity market value, which tend to have poor past returns, earn large excess
returns. It is also found that a similar relation exists between advertising and stock returns
and, R&D intensity is positively associated with returns volatility.
The study on investor psychology and asset pricing, Hirshleifer (2001) followed the
familiar quotations “the best plan is …to profit by the folly of others.” Further, the study
is influenced by Bill Blunte’s Deranged Anticipation and Perception Model (DAPM) in
which proxies for market misevaluation are used to predict security returns and concluded
that mispricing is only corrected slowly. The study is based on the survey and assesses the
theory and evidence regarding investor psychology as a determinant of asset pricing. In
the field of asset pricing, Campbell (2000) and Cochrane (2000) emphasize in external
sources of risk. The former argued that asset pricing is concerned with the sources of risk
and the economic forces that determine the rewards for bearing risk. And the later, stated
that the central task of financial economics is to figure out what are the real risks that
drive asset prices and expected returns. In contrast, the study argued that the central task
of asset pricing is to examine how expected returns are related to risk and to investor
misevaluation. The study concluded that the purely rational approach is being subsumed
by a broader approach based upon the psychology of investors. In this approach, security
expected returns are determined by both risk and misevaluation.
An analysis on market timing and capital structure is made by Baker and Wurgler (2002).
The findings of the study supported the generally accepted view that market timing is an
important aspect of real financing decisions. The study traced the implications of equity
market timing through to capital structure. Market-to-book ratio is used to measure the
market timing opportunities perceived by managers. The analysis found that low-leverage
firms tend to be those that raised funds when their valuations is high, and conversely
high-leverage firms tend to be those that raised funds when their valuations is low. The
study summarized that fluctuation in market valuations have large effects on capital
structure that persist for at least a decade. As a consequence, current capital structure is
strongly related to historical market values. It is well known that firms are more likely to
issue equity when their market values are high, relative to book value and past market
values, and to repurchase equity when their market values are low. In sum, the results
suggested the theory that capital structure is the cumulative outcome of past attempts to
time the equity market.
51
f) Review of major studies on market reactions to tangible and intangible
information
The information is broadly classified into the tangible and intangible components. Former
can be calculated from the firm’s financial statements and the later cannot be quantified in
numerical form. The accounting variables like the financial ratios, cash flows, sales, etc
are the examples of tangibles whereas the investor behavior, market behavior, news and
media impact, overconfidence, over and under-reactions, etc are the examples of
intangibles. The reviews of major studies in this section have been organized into five
sub-sections for tangible and intangible information.
i) Review of major studies on intangible information till 2000
The intangible issues like problems on risk measurements, financial ratings, trading halts,
the information diffusion process, etc and its effect on stock returns have been
incorporated into part. The study period covers 1989 to 2000 and the major findings of
the studies are presented in table 2.11 as follows:
Table 2.11: Review of major studies on market reactions to intangible information till 2000
Study Major findingsBernard and Thomas (1989)
Much of the evidences cannot plausibly be reconciled with arguments built on risk mismeasurement but is consistent with a delayed price response.
Goh and Ederington (1993)
Downgrades associated with deteriorating financial prospects convey new negative information to the capital market, but that downgrades due to changes in firms' leverage do not.
Lee, et.al (1994) Trading halts increase, rather than reduce, both volume and volatility. Brennan and Subrahmanyam (1995)
Other things equal, an increase in the number of investment analysts tends to be associated with reduction in the adverse selection cost of transacting.
Hong, et.al (2000) The study support Hong-Stein hypothesis which describes momentum reflects the gradual diffusion of firm-specific information. Thus, it is concluded that especially the negative information, diffuses only gradually across the investing public.
Bernard and Thomas (1989) attempted to discriminate between competing explanations
of post-earnings-announcement drift: a failure to adjust abnormal returns fully for risk
and a delay in the response to earnings reports. The empirical evidences showed that even
after earnings are announced, estimated cumulative abnormal returns continue to drift up
for good news firms and down for bad news firms. The study included the sample of
84,792 firm-quarters of data from NYSE and AMEX firms for the periods 1974-86 and
also conducted some supplementary tests based on 15,457 firm-quarters of data for over
the counter (OTC) stocks on the NASDAQ system from 1974 to 1985. Data were derived
52
from CRSP and COMPUSTAT database. Cumulative abnormal return (CAR) analysis as
a major statistical tool is employed in the study and concluded that much of the evidences
cannot plausibly be reconciled with arguments built on risk mismeasurement but is
consistent with a delayed price response.
In an effort to examine the stock returns and the bond ratings, Goh and Ederington (1993)
analyzed the reaction of common stock returns to bond rating changes. While the
financial literature exhibit a significant negative stock response to downgrades, the study
analyzes whether all downgrades are bad news for equity holders and whether all
downgrades are a surprise. When the rating agencies announce the rating changes, they
also have given the reasons. Based on these announced reasons, the study separate the
rating changes into groups based on whether they have positive or negative implications
for equity holders and whether or not they seem to be in response to recently released
public information. The study hypothesized that the negative reaction should not be
expected for all downgrades because: some rating changes are anticipated by market
participants; and the downgrades because of an anticipated move to transfer wealth from
bondholders to stockholders should be good news for stockholders. The study worked
with a set of 1078 rating changes announced by Moody's during the period 1984 through
1986, and excluded 468 because of insufficient data in CRSP daily returns database. The
study also searched the Wall Street Journal (WSJ) Index for other firm-specific
information releases in the three days surrounding the announcement date of the rating
change. If another announcement occurred during the three-day period, the rating change
announcement is eliminated yielding an uncontaminated sample of 428 ratings changes
(243 downgrades and 185 upgrades). Using Cumulative abnormal returns (CARs), the
study concluded that downgrades associated with deteriorating financial prospects convey
new negative information to the capital market, but that downgrades due to changes in
firms' leverage do not.
In the wake of the 1987 market break, a number of commentators on market mechanisms
recommended the establishment of "circuit breaker" mechanisms. The primary argument
supporting circuit breakers (both price limits and trading halts) is that non-trading periods
provide an opportunity for normal information transmission in times of market duress.
Proponents of circuit breakers claim that, during major price changes there can be a
breakdown in the transmission of information between the trading floor and market
participants. Therefore, "the primary function of a circuit breaker should be to re-inform
participants (Greenwald and Stein (1988)." Lee, et.al (1994) examined the volume,
53
volatility and trading halts in NYSE. The ISSM database is used for the analysis. The
initial sample consists of all trading halts on NYSE common stocks during 1988 as
identified by the ISSM database. Finally, a total of 42 observations were eliminated
yielding a sample of 852 trading halts. The study also look for the Wall Street Journal
Index, the New York Times Index, and the "Wires" and "Papers" services in the
Lexis/Nexis online database to classify the news events underlying the halts and used
only those articles that expressly mentioned the firm and provided a reason for the halt on
the day indicated by the ISSM tape. Then, classified the news into six general categories:
1) Acquisitions and Divestitures, 2) Capital Structure Changes, 3) Takeovers and
and 6) No News. With the mean abnormal volume and volatility statistics and the
regression analysis, the major findings of the study are: trading halts increase, rather than
reduce, both volume and volatility. Volume (volatility) in the first full trading day after a
trading halt is 230 percent (50 to 115percent) higher than following "pseudo-halts": non-
halt control periods matched on time of day, duration, and absolute net-of-market returns.
These results are robust over different halt types and news categories. Higher post-halt
volume is observed into the third day while higher post-halt volatility decays within
hours. The extent of media coverage is a partial determinant of volume and volatility
following both halts and pseudo-halts, but a separate halt effect remains after controlling
for the media effects.
Brennan and Subrahmanyam (1995) investigated the relation between the number of
analysis following a security and the estimated adverse selection cost of transacting in the
security, controlling for the effects of previously identified determinants. The study
defined the adverse selection costs of transacting as trading be investors who possess
superior information imposes significant liquidity costs on other market participants due
to adverse selection. Using intraday data of the year 1988, the study found that greater
analyst following tends to reduce adverse selection cost based on the market depth. In
other words, other things equal, an increase in the number of investment analysts tends to
be associated with reduction in the adverse selection cost of transacting.
Several recent studies have documented that, at medium-term horizons ranging from three
to 12 months, stock returns exhibit momentum – that is, past winners continue to perform
well, and past losers continue to perform poorly. Hong, et.al (2000) aimed to test the
Hong-Stein version of the underreaction hypothesis, namely: if momentum comes from
gradual information flow, then there should be more momentum in those stocks for which
54
information gets out more slowly. In other words, the study looks for evidence that
momentum reflects the gradual diffusion of firm-specific information. Using the data
from three primarily sources: the stock returns and turnover data from CRSP, the data on
analyst coverage are from the Institutional Brokers' Estimate System (IBES) and the
options-listing data come from the Options Clearing Corporation from 1976 to 1996 and
establish three key results. First, once one moves past the very smallest stocks, the
profitability of momentum strategies declines sharply with firm size. Second, holding size
fixed, momentum strategies work better among stocks with low analyst coverage. Finally,
the effect of analyst coverage is greater form stocks that are post losers than for past
winners. Thus, the study concluded that the findings are strongly consistent with the
Hong-Stein hypothesis, especially negative information, diffuses only gradually across
the investing public.
ii) Review of major studies on intangible information between 2000 and 2010
The internet message board activities, the effect of state of economy and the news, the
effect of information arrival to the public and its impact on stock returns, etc have been
managed in this sub-section. Table 2.12 presents the review of major studies and its major
findings as below:
Table 2.12: Review of major studies on market reactions to intangible information between 2000 to 2010
Study Major findingsTumarkin and Whitelaw (2001)
The internet message board activity did not predict industry-adjusted returns or abnormal trading volume, which is consistent with market efficiency.
Conrad, et.al (2002)
In explanation of the uncertainty about the state of the economy causes an asymmetry in the response to good news and bad news, the study support the hypothesis that stock prices respond most strongly to bad news in good times.
Vega (2006) There are not all information acquisition variables have the same effect on the market’s efficiency. Whether information is public or private is irrelevant; what matters is whether information is associated with the arrival rate of informed or uninformed traders.
Worthington (2006)
The study shows returns are highest during the ministries of Holt-McEwen and Hawke and lowest during Whitlam and Fraser, while risk is highest during Whitlam and Hawke and lowest during Menzies and Holt-McEwen. Thus, the study concluded that the risk differences potentially arise from the different parties’ economic and social policies, uncertainty among investors about these policies, or doubt among voters concerning future election outcomes.
Epstein and Schneider (2008)
Investors react asymmetrically to market signals as discount good news, but take bad news seriously.
Hertzberg, et.al (2010)
Loan Officers' rotation can be used to limit agency problems in communication and to detect the performance based bad news due to career concerns.
Financial economists believed that the internet is clearly playing an increasing role in
financial market and personal finance. The same issue has been analyzed as News or
55
Noise? internet posting and stock prices. Tumarkin and Whitelaw (2001) depicted that
there is growing effect of posting in internet financial forums affect stock prices, either
because the posting contain new information or because they represent successful
attempts to manipulate stock prices. The study examined the relationship between internet
message board activities and stock returns and trading volume during the study period
beginning from mid-April 1999 to mid-February 2000. The study focused on the
RagingBull.com discussion forum which is an extremely popular site whose format
permits the construction of an objective measure in investor opinions. The study found
that for stocks in the internet service sector, on days with abnormally high message
activities, changes in investor opinion correlated with abnormal industry-adjusted returns.
These event days also coincided with abnormally high trading volume, which persisted
for a second day. However, the study concluded that message board activity did not
predict industry-adjusted returns or abnormal trading volume, which is consistency with
market efficiency.
The growing concern in the capital market is, whether the price response to bad and good
earnings shocks changes as the relative level of the market changes. Conrad, et.al (2002)
examined this relationship. The study is based on the complete sample of annual earnings
announcements during the period 1988 to 1998, consensus earnings forecasts, realized
earnings, and earnings report dates are collected from IBES. The relative level of the
market is based on the difference between the current market price-earnings ratio and the
average market price-earnings ratio over the prior 12 months. The study explained that
the uncertainty about the state of the economy causes an asymmetry in the response to
good news and bad news. That is, when investors believe that the economy is in a bad
state and good news arrives, the inferred probability that the market is in a good state
increases; thus, the positive impact on prices is offset by the rising discount rate generated
by increased investor uncertainty. The aim of the study is to examine whether the strength
of firm-specific responses to new information is affected by the aggregate level of the
market. The study is worked against the a practitioner hypothesis quoted in the Wall
Street Journal, is that the stock prices of individual forms become relatively more
sensitive to bad news than good news as the market rises. This hypothesis is related to
two stands of literature. First, based on research in behavioral psychology, suggest that
investors inappropriately extrapolate past performance. Second, based on extended
regime shifting models, also predicts that the market will respond more strongly to bad
news than good news when stock prices are high. Then, the study suggested that both the
56
behavioral models and the regime shifting model could be extended to explain more fully
how the response of individual forms to earnings announcement may depend on the level
of the market. Primarily, the study found that the stock price response to negative
earnings surprises increases as the relative level of the market rises. Furthermore, the
difference between bad news and good news earnings response coefficients rise with the
market. In sum, the findings generally support the hypothesis that stock prices respond
most strongly to bad news in good times.
Worthington (2006) examined the presence of a political cycle in Australian daily stock
returns over 47 years from 6 January 1958 to 30 December 2005. The study period
includes 19 federal elections, 25 ministries and five terms of Liberal-National or Labor
government. The political cycle is defined in terms of the party in power, the time since
the last election and election information effects. The market variables are defined in
terms of nominal and real returns and nominal and real returns volatility. The results
indicates that highest returns during the ministries of Holt-McEwen and Hawke and
lowest during Whitlam and Fraser, while risk is highest during Whitlam and Hawke and
lowest during Menzies and Holt-McEwen. However, regression analysis shows that
Liberal-National and Labor governments more generally differ in the volatility of returns
where political cycle-sourced return volatility increases at a decreasing rate with the time
in power. Such risk differences potentially arise from the different parties’ economic and
social policies, uncertainty among investors about these policies, or doubt among voters
concerning future election outcomes. The study employs the parametric analysis to test
for a political cycle, a comparison of mean returns provides some empirical evidence to
support the conjecture that returns depend upon the government in power. There is
limited support for an election effect where returns are systematically higher or lower in
the period leading up to immediately following an election. Similar results are obtained
with a regression based analysis. Stock volatility reflects diffuse and easily changed
beliefs about future political behavior, but on balance, these views are never
systematically ‘bad’ or ‘good’ over extended periods of time.
In an efficient market, security prices at any given time fully reflect all available
information. Vega (2006) aimed to deepen the understanding on how private and public
information received by agents prior to earnings announcements affects the post-earnings
announcement drift. Data employed for the study are from six different sources: CRSP,
Compustat, ISSM, Trade and Quote (TAQ), IBES and Dow Jones Interactive. Finally, the
study obtained a final sample of 9,213 firms and 208,540 firm-quarter observations, and
57
used two variables to measure the amount of public information that is available to
investors prior to earnings announcements. Firstly, media coverage is defined as TAQ as
the number of days a particular firm is mentioned in the news prior to its earnings
announcement. Secondly, public news surprises is calculated when agents receive prior to
a firm’s earnings announcement using the stock market’s reaction to headline news. The
study used Easley and O’Hara (1992) information-based trading variables, personal
identification number (PIN), together with a comprehensive public news database to
empirically measure the effect of private and public information on the post-
announcement drift. The findings show that stocks associated with high PIN, consensus
public news surprises, and low media coverage experience low or insignificant drift.
Thus, there are not all information acquisition variables have the same effect on the
market’s efficiency. Whether information is public or private is irrelevant; what matters is
whether information is associated with the arrival rate of informed or uninformed traders.
Financial market participants absorb a large amount of news, or signals, every day.
Processing a signal involves quality judgments: News from a reliable source should lead
to more portfolios rebalancing than news from an obscure source. Unfortunately, judging
quality itself is sometimes difficult. It is true especially for tangible information, such as
earnings reports, that lends itself to quantitative analysis. By looking at past data,
investors may become quite confident about how well earnings forecast returns. Epstein
and Schneider (2008) focused on information processing when there is incomplete
knowledge about signal quality. The main idea is that, when quality is difficult to judge,
investors treat signals as ambiguous. In other words, the study analyzed the role of
uncertain information quality in financial markets. The proposed new model of
information processing by ambiguity-averse investors in the study, it is assumed that
investors perceive a range of signal precisions, and take a worst-case assessment of
precision when evaluating prior utility. The study documented that investors react
asymmetrically to signals: they discount good news, but take bad news seriously.
Moreover, they get disutility from low future information quality. The study also
emphasized three new effects of uncertain information quality on asset prices. First,
investors require compensation for low future information quality. Expected excess
returns are thus higher when information quality is more uncertain, holding fixed the
distribution of fundamentals. Second, investors require more compensation for low
information quality when fundamentals are more volatile. In markets in which
information quality is uncertain, expected excess returns thus scale with volatility, not
58
with covariance with the market or with marginal utility. Third, investors’ asymmetric
response to signals skews the distribution of observed returns: When there are signals of
uncertain quality, which generate negative skewness, signals of known quality generate
positive skewness. Thus, the study concluded that when ambiguity-averse investors
process news of uncertain quality, they act as if they take a worst-case assessment of
quality and react more strongly to bad news than to good news.
A rotation policy that routinely reassigns loan officers to borrowers of a commercial bank
affects the officers’ reporting behavior. When an officer anticipates rotation, reports are
more accurate and contain more bad news about the borrower’s repayment prospects.
Hertzberg, et.al (2010) presented the evidence that reassigning tasks among agents can
alleviate moral hazard in communication. As a result, the rotation policy makes bank
lending decisions more sensitive to officer reports. The threat of rotation improves
communication because self-reporting bad news has a smaller negative effect on an
officer’s career prospects than bad news exposed by a successor. Using data from the
internal records of the small and medium business division of the banks, the study
construct a monthly panel of loan officer-firm relationships. The sample covers the 7-year
period from December 1997, when the small and medium business division was created,
to December 2004. The observation includes 1,248 firms and 100 loan officers in 4,191
non-censored loan officer-firm relationships. In sum, the study concluded that rotation
can be used to limit agency problems in communication and to detect the performance
based bad news due to career concerns.
iii) Review of major studies on tangible information before 1990
Table 2.13 presents the review of major studies on tangible information before 1990s and
its major findings. Some of the key contributions during this period are: the positive
tradeoff between risk and returns, the effect of money supply in the stock returns, the
systematic price reversal, etc which has been organized as follows:
Table 2.13: Review of major studies on market reactions to tangible information before 1990
Study Major findingsFama and MacBeth (1973)
A positive tradeoff between return and risk.
Urich and Wachtel (1981)
The financial markets respond very quickly to the weekly money supply announcement.
De Bound and Thaler (1985)
Systematic price reversals for stocks that experience extreme long-term gains or losses: past losers significantly outperform past winners.
59
Cutler, et.al (1989)
The pieces of information, discount rates or the cash flows have large effects on stock prices.
An empirical work which tests the relationship between average return and risk for New
York Stock Exchange common stocks is performed by Fama and MacBeth (1973). The
theoretical basis of the tests is the "two-parameter" portfolio model and models of market
equilibrium derived from the two-parameter portfolio. The data for the study are monthly
percentage returns - including dividends and capital gains, and with the appropriate
adjustments for capital changes such as splits and stock dividends, for all common stocks
traded on the NYSE during the period January 1926 through June 1968. Related data
were derived from the CRSP. The sample period is divided into nine sub-periods for each
of portfolio formation, initial estimation period and for testing period. Rigorous
regression models are developed and tested in various forms, the behavior of the market
and components of the variance of the returns for the different sub-periods are analyzed in
varying forms of regression models. All above tests fail to reject the hypothesis of these
models that the pricing of common stocks reflects the attempts of risk-averse investors to
hold portfolios that are "efficient" in terms of expected value (return), and dispersion of
return (risk). Moreover, the observed "fair game" properties of the coefficients and
residuals of the risk-return regressions are consistent with an efficient capital market- that
is, a market where prices of securities fully reflect the available information. In
conclusion, the results support the important testable implications of the two parameter
model. Given that the market portfolio is efficient or, more specifically, given that proxy
for the market portfolio is at least approximately efficient. Thus, the tests cannot reject
the hypothesis that average returns on the NYSE common stocks reflect the attempts of
risk-averse investors to hold efficient portfolios. Specifically, on average there seems to
be a positive tradeoff between return and risk. Thus, the evidence support the hypothesis
that in making a portfolio decision, an investor should assume that the relationship
between a security's portfolio risk and its expected return is linear, as implied by the two-
parameter model. Also, the study cannot reject the hypothesis of the two-parameter model
that no measure of risk, in addition to portfolio risk, systematically affects average
returns. Finally, the observed fair game properties of the coefficients and residuals of the
risk-return regressions are consistent with an efficient capital market.
The predicted values from ARIMA models and the median forecasted change from the
survey are used as the expected money supply in a study of macroeconomic effects on
stock returns. Urich and Wachtel (1981) analyzed the market response to the weekly
60
money supply announcements. Interest rates, money supply and the expected money
supply are three variables of the analysis. The period under consideration for the study is
1970 to 1979. The major finding of the study is the financial markets respond very
quickly to the weekly money supply announcement.
Do the Stock Market Overreact? The motivation for De Bound and Thaler (1985) is the
other research in experimental psychology suggests that, in violation of Bayes' rule, most
people tend to overreact to unexpected and dramatic news events. What is an appropriate
reaction? Both classes of behavior that is market behavior and the psychology of
individual decision making behavior can be characterized as displaying overreaction. For
example, individuals tend to overweight recent information and underweight prior and in
spite of the observed trendiness of dividends, investors seem to attach disproportionate
importance to short-run economic developments. Amid these observations, the study is
undertaken to investigate the possibility that the market behavior and the psychology of
individual decision making are related by more than just appearance. In other words, the
goal is to test whether the overreaction hypothesis is predictive i.e. whether such behavior
affects the stock prices. The findings are based on NYSE common stocks monthly return
data retrieved from CRSP for the period 1926 to 1982, is consistent with the overreaction
hypothesis. Specifically, two more hypotheses are tested; first, extreme movements in
stock prices will be followed by subsequent price movements in the opposite direction
and the other, more extreme the initial price movement, the greater will be the subsequent
adjustment. Both hypotheses imply a violation of weak-form of market efficiency.
Substantial weak form of market inefficiencies is discovered. The results also shed new
light on the January returns earned by prior winners and losers. Portfolios of losers
experience exceptionally large January returns as late as five years after portfolio
formation which is concluded with the analysis of cumulative average residuals for
winner and loser portfolios. Portfolios are formed of the 50 most extreme winners and 50
most extreme losers measured by cumulative excess returns over successive five year
formation periods. Thus, the major outcome of the study is systematic price reversals for
stocks that experience extreme long-term gains or losses: Past losers significantly
outperform past winners.
Several studies of asset pricing have challenged the view that stock price movements are
wholly attributable to the arrival of news. Cutler, et.al (1989) estimated the fraction of the
variance in aggregate stock returns that can be attributed to various types of economic
news. To understand whether unexpected macroeconomic developments can explain a
61
significant fraction of share price movements, the study analyzes monthly stock returns
for the 1926-1985 period, as well as annual returns for the longer 1971-1986 period.
Regression models and less structured approaches are used to examine the news and the
study found that the arrival of information about macroeconomic performance news
proxies can explain about one-third of the variance in stock returns; while analyzing the
stock market reactions to identifiable world news – regarding wars, the Presidency, or
significance changes in financial policies affect stock prices, the findings cast doubt on
the view that qualitative news can account for all the return variation that cannot be traced
to macroeconomic innovations. Thus, the study argued that further understanding of asset
price movements requires two types of research. The first should attempt to model price
movements as functions of evolving consensus opinions about the implications of given
pieces of information. The second should develop and test ‘propagation mechanisms’ that
can explain why stocks with small effects on discount rates or cash flows may have large
effects on prices.
iv) Review of major studies on tangible information during 1990s
The tangible information for instance, macroeconomic indicators, the intraday trading
patterns, the effect of dividend omission, etc during the period 1990s have been organized
in this sub-section. The major findings of such studies have been presented in Table 2.14
as follows.
Table 2.14: Review of major studies on market reactions to tangible information during 1990s
Study Major findingsEderington and Lee (1993)
The scheduled macroeconomic news announcements are responsible for most of the observed time-of-day and day-of-week volatility patterns. Thus, there is greatest impact of these announcements on interest rate and foreign exchange futures markets.
Berry and Howe (1994)
The public information arrival is nonconstant, displaying seasonalities and distinct intraday patterns - information arrival exhibits an inverted U-shape pattern across trading days. Next, the study relate the measure of public information to aggregate measures of intraday market activity which suggest a positive, moderate relationship between public information and trading volume, but an insignificant relationship with price volatility.
Michaely, et.al (1995)
The magnitudes of short-run price reactions to dividend omissions are greater than for dividend initiations. In the year following the announcements, prices continue to drift in the same direction, though the drift following omissions is stronger and more robust.
Braun, et.al (1995)
There is strong evidence of conditional heteroskedasticity in both market and non-market components of returns, and weaker evidence of time-varying conditional betas.
La Porta, et.al (1997)
The announcement returns suggest a significant portion of the return difference between value and glamour stocks is attributable to earnings surprises that are systematically more positive for value stocks.
62
Impact of scheduled macroeconomic news announcements on interest rate and foreign
exchange futures markets is examined by Ederington and Lee (1993). The study found
these announcements are responsible for most of the observed time-of-day and day-of-
week volatility patterns in these markets. While the bulk of the price adjustment to a
major announcement occurs within the first minute, volatility remains substantially higher
than normal for roughly fifteen minutes and slightly elevated for several hours.
Nonetheless, these subsequent price adjustments are basically independent of the first
minute’s returns. Thus, the study identified those announcements with the greatest impact
on interest rate and foreign exchange future markets.
The link between information and changes in asset prices is the central issue in financial
economics. A fundamental tenet of market efficiency is that investors react to new
information as it arrives, resulting in price changes that reflect investors' expectations of
risk and return. Berry and Howe (1994) developed a measure of public information flow
to financial markets and used it to document the patterns of information arrival, with an
emphasis on the intraday flows. The measure is the number of news releases by Reuter's
News Service per unit of time. Over 120,000 observations were collected from May 1990
to April 1991. The database contained all information events, not only firm-specific
information, over the full 24-hour day. The study found the public information arrival is
non-constant, displaying seasonality and distinct intraday patterns - information arrival
exhibits an inverted U-shape pattern across trading days. Next, the study relate the
measure of public information to aggregate measures of intraday market activity which
suggest a positive, moderate relationship between public information and trading volume,
but an insignificant relationship with price volatility.
When a firm initiates the payment of a cash dividend, or omits such a payment, the firm is
making an extremely visible and qualitative change in corporate policy. What effect do
such abrupt changes have on returns? Michaely, et.al (1995) investigates both the
immediate (three-day) reaction to initiation or omission announcements and the long-term
post-announcement price performance. The study used CRSP tapes to collect all NYSE
and AMEX companies that initiated dividends during 1964 to 1988 and defined a
dividend initiation as the first cash dividend payment reported on the CRSP master file,
reinstitution of a cash dividend is not considered. During 25 years, total 561 cash
dividend events are considered the resulting sample for the study. Consistent with prior
literature, the study concluded that the magnitudes of short-run price reactions to
omissions are greater than for initiations. In the year following the announcements, prices
63
continue to drift in the same direction, though the drift following omissions is stronger
and more robust.
Many studies have been documented that stock returns volatility tend to rise following
good and bad news, this phenomenon, which Braun, et.al (1995) explained as predictive
asymmetry of second moments. The study investigated the conditional covariance of
stock returns using bivariate exponential ARCH (EGARCH) models. These models allow
market volatility, portfolio specific volatility, and beta to respond asymmetrically to
positive and negative market and portfolio returns, i.e., "leverage" effects. Using monthly
data, the study depicted strong evidence of conditional heteroskedasticity in both market
and non-market components of returns, and weaker evidence of time-varying conditional
betas. Surprisingly while leverage effects appear strong in the market component of
volatility, they are absent in conditional betas and weak and/or inconsistent in nonmarket
sources of risk.
Most of the financial researches agreed that simple value strategies based on such ratios
as book-to-market, earnings-to-price and cash flow-to-price have produced superior
returns over a long period of time. La Porta, et.al (1997) examined the hypothesis that the
superior return to so-called value stocks is the result of expectational errors made by
investors. The study examined the stock price reactions around earnings announcements
for value and glamour stocks over a 5-year period after portfolio formation. The study
used the sample firms of NYSE, AMEX and Nasdaq that are available on Compustat and
CRSP tapes. The sample period runs from 1971:2 through 1993:1. To examine earnings
announcement return differences between value and glamour stocks, the study form
portfolios on the basis of two classifications: the book-to-market ratio and two-way
classification based on cash-flow-to-price and past growth-in-sales. The findings are:
expectational errors about future earnings prospects play an important role in the superior
return to value stocks, post-formation earnings announcement returns are substantially
higher for value stocks than for glamour stocks, event returns for glamour stocks are
significantly lower than glamour returns on an average day, which is inconsistent with the
risk premium explanation for the return differences between value and glamour stocks,
and in full sample, earnings announcement return differences account for approximately
25-30 percent of the annual return differences between value and glamour stocks in the
first two to three years after portfolio formation and approximately 15-20 percent of
return differences over years four and five after formation. Thus, the study concluded that
the announcement returns suggest a significant portion of the return difference between
64
value and glamour stocks is attributable to earnings surprises that are systematically more
positive for value stocks. The evidence is inconsistent with a risk-based explanation for
the return differential.
v) Major studies on tangible information 2000 onwards
Table 2.15 shows the review of major studies and its key findings regarding market
reactions to tangible information for the period 2000 onwards are as follows.
Table 2.15: Review of major studies on market reactions to tangible information 2000 onwards
Study Major findingsVeronesi (2000) The relationship between the precision of public information about economic growth
and performance of the stock market is nontrivial.
Andersen, et.al (2000)
The intra-daily volatility exhibits a doubly U-shaped pattern associated with the opening and closing of the separate morning and afternoon trading sessions on the Tokyo Stock Exchange, which is consistent with market microstructure theories that emphasize the role of private and asymmetric information in the price formation process.
Brav and Lehavy (2003)
There is a significant market reaction to the information contained in analysts’ target prices, both unconditionally and conditional on contemporaneously issued stock recommendation and earnings forecast revisions.
Grinblatt and Moskowitz (2004)
The consistency of positive past returns and tax-loss selling significantly affects the relation between past returns and the cross-section of expected returns.
Domer (2005) Public financial information has an impact on stock market behavior. Thus, the study concluded a positive correlation between the stock prices and the information categories: net asset value, occupancy rates, cash flow and overall capitalization rate.
Kothari, et.al (2006)
Market reaction to aggregate earnings is much different than the reaction to firm earnings and there is little evidence that prices react slowly to aggregate earnings news and the behavioral theories that explain post-earnings announcement drift in firm returns do not seem to describe aggregate price behavior.
Watanabe (2008) Accurate information increases the volatility. Eaves and Williams (2010)
The intraday volume is U-shaped, intraday volatility is closer to L-shaped even if the previous studies reported U-shaped intraday volatility, and the study concluded that the timing of privately informed traders cannot be the source of intraday patterns.
In modern financial markets, investors are flooded with a variety of information: firms’
earnings reports, revisions of macroeconomic indexes, policymakers’ statements, and
political news. These pieces of information are processed by investors to update their
projections of the economy’s future growth rate, inflation rate, and interest rate. Veronesi
(2000) using a simple dynamic asset pricing model, investigated the relationship between
the precision of public information about economic growth and stock market returns.
After fully characterizing expected returns and conditional volatility, the study
documented that higher precision of signals tends to increase the risk premium; when
signals are imprecise or noisy the equity premium is bounded above independently of
investors’ risk aversion; return volatility is U-shaped with respect to investors’ risk
65
aversion; and the relationship between conditional expected returns and conditional
variance is ambiguous. Thus, the study showed that the relationship between the precision
of public information about economic growth and performance of the stock market is
nontrivial.
The volatility in the Japanese stock market based on a 4-year sample of 5-min Nikkei 225
returns from 1994 through 1997, is characterized by Andersen, et.al (2000). The intra-
daily volatility exhibits a doubly U-shaped pattern associated with the opening and
closing of the separate morning and afternoon trading sessions on the Tokyo Stock
Exchange. This feature is consistent with market microstructure theories that emphasize
the role of private and asymmetric information in the price formation process.
Meanwhile, readily identifiable Japanese macroeconomic news announcements explain
little of the day-to-day variation in the volatility, confirming previous findings for US
equity markets. Furthermore, by appropriately filtering out the strong intraday periodic
pattern, the high-frequency returns revealed the existence of important long-memory
intra-daily volatility dependencies.
In recent years, security analysts have been increasingly disclosing target prices, along
with their stock recommendations and earnings forecasts. Using a large database of
analyst price targets, stock recommendations, and earnings forecasts, Brav and Lehavy
(2003) examined the short-term market reactions to target price announcements and long-
term co-movement of target and market prices. Using a large database of analysts’ target
prices issued over the period 1997 to1999, the study found a significant market reaction
to the information contained in analysts’ target prices, both unconditionally and
conditional on contemporaneously issued stock recommendation and earnings forecast
revisions. Using a co-integration approach, the study analyzes the long-term behavior of
market and target prices and found that, on average, the one-year-ahead target price is 28
percent higher than the current market price. Moreover, revisions in target prices contain
information about six-month post-event abnormal returns. Recommendation and earnings
forecast revisions are also found to be informative in the presence of target prices. Since,
the study aimed to explore and document the evidence on the informativeness and time-
series behavior of analysts’ target prices, contributes to the understanding of price
formation in equity markets.
Grinblatt and Moskowitz (2004) analyzed the consistency and sign of the past return as
well as the degree to which tax-motivated trading generates effects on future returns. Both
66
short-term and long-term past returns contain information about expected returns because
they are the proxy for a more fundamental variable that predicts returns. A key finding of
the analysis is that winner consistency is important. The consistency can also be the proxy
for inverse of volatility and this may affect average returns as a proxy for risk. Achieving
a high past return with a series of steady positive months appears to generate a larger
expected return than a high past return achieved with just a few extraordinary months.
The study also highlighted the importance of seasonality associated with past returns and
the degree to which tax-loss trading plays a role in past return predictability. The study
used the approach in analyzing the importance of these complex patterns of returns is
based on a parsimonious stock ranking system derived from simple Fama-MacBeth cross-
sectional regressions. The study analyzes the simultaneous effect of a number of past
return-related variables on the future returns of hedged positions in individual stocks,
which have their size, book-to-market, and industry return components eliminated and the
beta neutral as well. The sample constituted the monthly returns from every records of
security on the CRSP data files from August 1963 to December 1999. From 1963 to
1973, the CRSP sample includes NYSE and AMEX firms only, and post-1973
NASDAQ-NMS firms are added to the sample. Thus, the consistency of positive past
returns and tax-loss selling significantly affects the relation between past returns and the
cross-section of expected returns. Analysis of these additional effects across stock
characteristics, seasons, and tax regimes provides clues about the sources of temporal
relations in stock returns, pointing to potential explanations for this relation. A
parsimonious trading rule generates surprisingly large economic returns despite controls
for confounding sources of return premia, microstructure effects, and data snooping
biases.
The analysis on responses of stock prices to financial announcements is examined by
Domer (2005). The study employed a computer-based content analysis of qualitative data.
The data is from a Swedish real estate firm during the period 1991-1996. The information
collected and analyzed comes from the company's press releases, quarterly statements and
articles in the six largest business magazines in Sweden. Annual statements were
excluded from the study due to the fact that the publication of these statements occurred
at a point of time when the information was already reflected in the share prices. The
major finding of the study is a positive correlation between the stock prices and the
following information categories: net asset value, occupancy rates, cash flow and overall
capitalization rate. These results are compared to other studies investigating the influence
67
of information on stock prices. The results of the study confirmed previous results. Thus,
the main contribution of the study is that it supports the assumption that public financial
information has an impact on stock market behavior.
A study on stock market's reaction to aggregate earnings news, Kothari, et.al (2006)
found a substantially different pattern in aggregate earning and returns. In prior studies,
for individual firms, stock prices react positively to earnings news but require several
quarters to fully reflect the information in earnings. The study concluded two major
findings: First, returns are unrelated to past earnings, suggesting that prices neither
underreact nor overreact to aggregate earnings news. Second, aggregate returns correlate
negatively with concurrent earnings during the study period 1970 to 2000. The earnings
series include all NYSE, AMEX, and NASDAQ stocks with data for earnings, price, and
book equity on the Compustat Quarterly file from 1970 – 2000. The market return is the
CRSP value-weighted index and compound monthly index returns to obtain quarterly
returns. Fama-MacBeth regression, time series analysis, correlation, auto-regressive
models, autocorrelations and behavioral models were used for the analysis. The study
also suggests that earnings and discount rates move together over time which is
inconsistent with asset-pricing models that imply discount rates and cash flows move in
opposite directions, and provides new evidence that discount-rate shocks explain a
significant fraction of aggregate stock returns. In conclusion, the market’s reaction to
aggregate earnings is much different than the reaction to firm earnings and there is little
evidence that prices react slowly to aggregate earnings news and the behavioral theories
that explain post-earnings announcement drift in firm returns do not seem to describe
aggregate price behavior.
A study on price volatility and investor behavior in an Overlapping Generations Model
with information asymmetry is conducted by Watanabe (2008). The study begins with the
issue: the mounting evidence of both trend-following and contrarian behavior among
various investor groups in recent empirical studies. Trend-followers buy assets upon price
appreciation and sell them upon depreciation, while contrarians trade in the opposite way.
Such trading behavior is found in both domestic and international markets. Moreover,
prices in these markets are found to vary much more than the stocks’ fundamental values.
The study follows an Overlapping Generations Model (OGM) with multiple securities
and heterogeneously informed agents. The study first found that asset prices can be highly
volatile relative to dividend variability and less informed agents rationally behave like
trend-followers, while better informed agents follow contrarian strategies. Trading
68
volume has a hump-shaped relation with information precision and is positively
correlated with absolute price changes. Calibrating the Full-Information Model and
explaining the phenomena from a fully rational perspective, the empirical studies
documented that various investor classes follow trend-chasing and contrarian strategies in
both domestic and international markets. Many of these markets are found to exhibit
excess volatility and, in some cases, strong co-movements in asset returns. Finally,
accurate information increases the volatility and correlation of stock returns in highly
volatile and strongly correlated equilibrium.
In a study of intraday volume and volatility, Eaves and Williams (2010) documented that
no matter how pronounced intraday patterns may appears, it is difficult to account for
cross-correlations among related assets when those assets trade continuously and
simultaneously. Based on the practice, futures contracts are auctioned periodically and
sequentially on the Tokyo Grain Exchange (TGE) the study analyses the intraday volume
and volatility. The dataset covers the 1,407 business days from May 1994 through
January 2000, encompassing 5,540 trading sessions for corn, 8,394 sessions for red beans,
5,596 for soybeans, and 7,004 for sugar. Even though intraday TGE volume is U-shaped,
intraday volatility is closer to L-shaped even if the previous studies reported U-shaped
intraday volatility for example the market microstructure theory which sought to explain
why intraday volatility is U-shaped. After accounting for the public information in
immediately preceding auctions for the same commodity, for earlier trading in other
commodities, and for trading on overseas markets open overnight in Tokyo, the intraday
patterns are effectively flat. Thus, the timing of privately informed traders cannot be the
source of intraday patterns.
g) Review of major studies related to media effects
The media coverage, public relations, other investor behavior and stock returns, media
optimism and pessimism and its relation to stock returns, high and low media coverage
and stock returns, etc are reviewed in this sub-section. The study period covered 1987 to
2011. Table 2.16 shows the studies and its major findings as follow.
The Efficient Market Hypothesis (EMH) assumes that the real-world investors at the time
of their portfolio decisions have access to the complete prior history of all stock returns.
When, however, investors’ decisions are made, the price data may not have been in
reasonably-accessible form and the computational technology necessary to analyze all
these data may not even have been invented. In such cases, the classification of all prior
69
price data as part of the publicly available information set may introduce an important
bias. The study developed a two-period model of capital market equilibrium in an
environment where each investor knows only about a subset of the available information.
Thus, the study analyzes the impact on the structure of equilibrium asset prices caused by
the particular type of incomplete information. The comparative statistics are used to
Table 2.16: Review of major studies on media effect on stock returnsStudy Major findingsMerton (1987) Financial markets dominated by rational agents may nevertheless produce anomalous
behavior relative to the perfect-market model. Thus, media coverage, public relations and other investor marketing activities could play an important causal role in creating and sustaining speculative bubbles and fads among investors.
Tetlock (2007) The study primarily contributed the three things as: First and foremost, the study found that high levels of media pessimism robustly predict downward pressure on market prices, followed by a reversion to fundamentals. Second, unusually high or low values of media pessimism forecast high market trading volume. Third, low market returns lead to high media pessimism.
Fang and Peress (2009)
High-media coverage stocks earn lower returns.
Engelberg and Parsons (2011)
The presence or absence of local media coverage is strongly related to the probability and magnitude of local trading.
analyze the cross-sectional differences among expected returns. Total of 1387 sample
firms were taken from the COMPUSTAT tapes as on December 31, 1985. Merton (1987)
found that media coverage, public relations and other investor marketing activities could
play an important causal role in creating and sustaining speculative bubbles and fads
among investors, the expanded media coverage of a firm, industry or other sector of the
economy is stimulated by changes in the same economic fundamentals that cause firms to
change their plans and investors to reassess their portfolio, advertising that initially
attracts investor attention to a firm is assumed to leave that firm’s investor base
unchanged if the underlying fundamentals do not justify a change. Thus, the study
concluded that financial markets dominated by rational agents may nevertheless produce
anomalous behavior relative to the perfect-market model. Institutional complexities and
information costs may cause considerable variations in the time scales over which
different types of anomalies are expected to be eliminated in the market place. Whether or
not the specific information inefficiency posited can be sustained in the long run, the
model may nevertheless provide some intermediate insights into the behavior of security
prices.
Causal observation suggests that the content of news about the stock market could be
linked to investor psychology and sociology. However, it is unclear whether the financial
news media induces, amplifies, or simply reflects investors’ interpretations of stock
70
market performance. Tetlock (2007) attempted to characterize the relationship between
the content of media reports and daily stock market activity, focusing on the immediate
influence of the Wall Street Journal’s “Abreast of the Market” column on U.S. stock
market returns over the 16 year period from 1984 to 1999 and 7895 words and verbs were
analyzed and classified into 7 broad groups and 77 categories based on Harvard
dictionary. Using principal components analysis (PCA), the study construct a simple
measure of media pessimism from the content of the WSJ column, then estimate the inter-
temporal links between this measure of media pessimism and the stock market using
basic vector autoregressions. First and foremost, the study found that high levels of media
pessimism robustly predict downward pressure on market prices, followed by a reversion
to fundamentals. Second, unusually high or low values of media pessimism forecast high
market trading volume. Third, low market returns lead to high media pessimism. These
findings suggested that measures of media content serve as a proxy for investor sentiment
or non-informational trading.
On a study of media coverage and the cross-section of stock returns, Fang and Peress
(2009) tested the hypothesis, by reaching a broad population of investors, mass media can
alleviate informational frictions and affect security pricing even if it does not supply
genuine news. The study investigates this hypothesis by studying the cross-sectional
relation between media coverage and expected stock returns and found that stocks with
no media coverage earn higher returns than stocks with high media coverage even after
controlling for well-known risk factors. These results are more pronounced among small
stocks and stocks with high individual ownership, low analyst following, and high
idiosyncratic volatility. The findings suggest that the breadth of information
dissemination affects stock returns. The sample considered all the listed companied on the
NYSE, contains mainly large stocks and 500 randomly selected companies listed on the
NASDAQ between 1993 and 2002. Univariate analysis, comparison of average returns of
stocks with firm characteristics and media coverage and multivariate analysis, four
different factor models: the market model, the Fama-French (1993) three-factor model,
the Carhart (1997) four-factor model, and a five-factor model that includes the Pastor-
Stambaugh (2003) liquidity factor has used for the analysis. The major finding of the
study is that high-media coverage stocks earn lower returns.
On a study of media and financial markets, Engelberg and Parsons (2011) analyzed the
causal impact of media in financial markets. The objective is to disentangling the causal
impact of media reporting from the impact of the events being reported. The study is
71
conducted by comparing the behaviors of investors with access to different media
coverage of the same information event. The study followed two approaches: the first is
to select the events which is the determinants of media coverage and market responses
can be decoupled, Brute-force approach. The second is cross-sectional approach, the basic
idea of this is to take two groups of agents and for the same information event, vary only
media exposure. This study is primarily focus on second approach. Using the multivariate
regression model for 19 mutually exclusive trading regions corresponding with large U.S.
cities, the study found that local media coverage strongly predicts local trading, after
controlling for earnings, investor, and newspaper characteristics for all earnings
announcements of S&P 500 Index firms. Moreover, local trading is strongly related to the
timing of local reporting. Thus, analyzing the simultaneous reactions of investors in 19
local markets to the same set of information events like earnings releases of S&P 500
Index firms, the study concluded that the presence or absence of local media coverage is
strongly related to the probability and magnitude of local trading.
h) Review of major studies on news effects
The number of news stories and market activities might not be associated, the news
events like: dividends disclosure, bonus and right announcements, financial disclosure,
etc might have effect on stock returns. But, what are the evidences available for these
variables are presented in this sub-section which includes the major studies during 1992
to 2009. The review of major studies during the period and its key contributions are
presented in Table 2.17 as follows:
One striking characteristic of the stock market is that the volatility of returns can be very
different at different times; daily volatility also fluctuates, and can change very rapidly. It
seems plausible that changes in volatility may have important effects on required stock
returns, and thus on the level of stock prices. Campbell and Hentschel (1992) analyzed
the volatility feedback mechanism or ‘no news is good news’ by modifying the GARCH
model of returns to allow for volatility feedback effect. The study emphasized all large
pieces of news have a negative volatility effect; conversely, all small pieces of news have
a positive volatility effect. The arrival of a small piece of news lowers future expected
Table 2.17: Review of major studies on news effect on stock returnsStudy Major findingsCampbell and Hentschel (1992)
The study concluded that volatility feedback contributes little to the unconditional variance of returns thus much of the variance of stock is in fact due to other changes in expected excess returns, and not to news about future dividends.
Mitchell and The study concluded the number of news stories and market activity are directly related.
72
Mulherin (1994) Maheu and McCurdy (2004)
The study interprets the innovation to returns, which is directly measurable from price data, as the news impact from latent news innovations. The latent news process is postulated to have two separate components, normal news and unusual news events, which have different impacts on returns and expected volatility for individual stocks.
Boyd, et.al (2005)
On average, the stock market responds positively to news of rising unemployment in expansions and negatively in contractions. Since, the economy is usually in an expansion phase, it follows that the stock market usually rises on the announcement of bad news from the labor market.
Zhang (2006) Greater information uncertainty produce relatively higher expected returns following good news and relatively lower expected returns following bad news.
Hirshleifer, et.al (2009)
The univariate and multivariate tests provide statistically significant evidence that high-news days are associated with a lower sensitivity of announcement abnormal returns to earnings news, a higher sensitivity of post-announcement abnormal returns to earnings news, and a lower trading volume response to earnings news.
volatility and increases the stock price. In the extreme case where no news arrives, the
market raises because ‘no news is good news’. Volatility feedback therefore implies that
stock price movements will be correlated with future volatility. The resulting model is
asymmetric because volatility feedback amplifies large negative stock returns and
dampens large positive returns. The model also implies that volatility feedback is more
important when volatility is high. In US monthly and daily data in the period 1926-1988,
the asymmetric model fits the data better than the standard GARCH model, accounting
for almost half the skewness and excess kurtosis of standard monthly GARCH residuals.
Estimated volatility discounts on the stock market range from 1 percent in normal times
to 13 percent after the stock market crash of October 1987 and 25 percent in early 1930s.
However volatility feedback has little effect on the unconditional variance of stock
returns. The study also explained that the basic problem is that stock returns are
determined endogenously in general equilibrium, one cannot explain the behavior of
stock returns in economic terms by applying a statistical model directly to returns; and it
is further feature that volatility feedback is more important when volatility is high than
when volatility is low. In sum, the study concluded that volatility feedback contributes
little to the unconditional variance of returns thus much of the variance of stock is in fact
due to other changes in expected excess returns, and not to news about future dividends.
Whether the amount of information that is publicly reported affects the trading activity
and the price movements in securities markets is the capital market issue. Mitchell and
Mulherin (1994) focused on the relationship between the numbers of news
announcements reported daily by Dow Jones & Company and aggregate measures of
securities market activity including trading volume and market returns. The study found
73
that the number of Dow Jones announcements and market activity are directly related.
The results are robust to the addition of factors previously found to influence financial
markets such as day-of-the-week dummy variables, news importance as proxy by large
New York Times headlines and major macroeconomic announcements, and non-
information sources of market activity as measured by dividend capture and triple
witching trading. However, the observed relationship between news and market activity is
not particularly strong and the patterns in news announcements do not explain the day-of-
the-week seasonality in market activity. The analysis of the Dow Jones database
confirmed the difficulty of linking volume and volatility to observed measures of
information. The data cover 2,011 business days during 1983 to 1990. Using the time
series pattern analysis, correlation and the regression, the study concluded that the
number of news stories and market activity are directly related.
There is a wide-spread perception in the financial press that volatility of asset returns has
been changing. The new economy is introducing more uncertainty. Indeed, it can be
argued that volatility is being transferred from the economy at large into the financial
markets, which bear the necessary adjustment shocks. In order to assess the empirical
validity of the perception and to investigate the sources and characteristics of changing
volatility dynamics on many important financial and economic decisions, Maheu and
McCurdy (2004) modeled the components of the return distribution, which are assumed
to be directed by a latent news process. The study interprets the innovation to returns,
which is directly measurable from price data, as the news impact from latent news
innovations. The latent news process is postulated to have two separate components,
normal news and unusual news events, which have different impacts on returns and
expected volatility for individual stocks. Normal news innovations are assumed to cause
smoothly evolving changes in the conditional variance of returns where as the unusual
news process causes infrequent large moves in returns. The study selected the random
sample of 11 US firms, daily price data for the randomly chosen firms that fit the sample
criteria were obtained from the CRSP database at the end of December, 2000 and used
three indices – DJIA, Nasdaq 100 and CBOE Technology Index (TXX). GARCH-jump
model with autoregressive jump intensity (GARJI) is applied for the analysis. The
conditional variance of stock returns is a combination of jumps and smoothly changing
components. A heterogeneous Poisson process with a time-varying conditional intensity
parameter governs the likelihood of jumps. Unlike typical jump models with stochastic
volatility, previous realizations of both jump and normal innovations can feedback
74
asymmetrically into expected volatility. The study developed the model which improve
the forecasts of volatility, particularly after large changes in stock returns and provide the
empirical evidence of the impact and feedback effects of jumps versus normal return
innovations, leverage effects, and the time-series dynamics of jump clustering.
In short-run response of stock prices to the arrival of macroeconomic news – the
announcement of the unemployment rate, Boyd, et.al (2005) presented the stock market’s
response to unemployment news arrival depends on whether the economy is expanding or
contracting. On average, the stock market responds positively to news of rising
unemployment in expansions and negatively in contractions. Since, the economy is
usually in an expansion phase, it follows that the stock market usually rises on the
announcement of bad news from the labor market. The monthly unemployment
announcements database are used for the study and it cover the period from February
1957 to December, 2000. The regression models are used to analyze the stock and bond
price responses to unemployment news. The explanation of the findings - the seemingly
odd pattern of stock price responses are: Conceptually, three primitive factors determine
stock prices – the risk-free rate of interest, the expected rate of growth of earnings and
dividends or the growth expectations, and the equity risk premium. Thus, if
unemployment news has an effect on stock prices, that must be because it conveys
information about one or more of these primitives.
The study documented the relationship between information uncertainty and stock
returns. Based on the substantial evidence of short-term stock price continuation, the prior
literature often attributes to investor behavioral biases such as underreaction to new
information. Zhang (2006) investigated the role of information uncertainty in price
continuation anomalies and cross-sectional variations in stock returns using sample data
from three sources. Returns from the CRSP monthly stocks which include NYSE,
AMEX, and NASDAQ stocks, book value and other financial data are from Compustat,
analyst forecast revisions are from IBES. The sample period spans from January 1983 to
December 2001. Since, Fama and French three-factor model does not capture the
momentum effect, the study used a four-factor model to test portfolio returns. The general
market reaction principle is good news predicts relatively higher future returns and bad
news predicts relatively lower future returns. If short-term price continuation is due to
investor behavioral biases, there should be greater price drift when there is greater
information uncertainty. Specifically, the study focused on two price continuation
anomalies: post-analyst forecast revision price drift and price momentum, using ex-post
75
returns as a proxy for expected returns, the analysis found consistent results across six
proxies for information uncertainty: firm size, firm age, analyst coverage, dispersion in
analyst forecasts, return volatility, and cash flow volatility. For each of the six proxies,
greater information uncertainty leads to relatively lower future stock returns following
bad news and relatively higher future returns following good news, suggested that
uncertainty delays the flow of information into stock prices. In sum, the study concluded
that greater information uncertainty should produce relatively higher expected returns
following good news and relatively lower expected returns following bad news.
Recent financial literature proposes that limited investor attention causes market
underreactions, the explanation of underreaction is that investors with limited attention
neglect newly arriving information signals. Hirshleifer, et.al (2009) examined this
explanation by measuring the information load faced by investors. The study provides
new insight into the validity of the attention hypothesis by testing directly whether
extraneous news distracts investors, causing market prices to underreact to relevant news.
The investor distraction hypothesis, which holds that the arrival of extraneous earnings
news causes trading volume and market prices to react sluggishly to relevant news about
a firm. Specifically, the study examined how the number of earnings announcements by
other firms affects a firm’s volume, announcement period return, and post-event return
reactions to an earnings surprise. Using the quarterly earnings announcement data from
the CRSP-Compustat merged database and IBES from 1995 to 2004, the study found that
the immediate price and volume reaction to a firm’s earnings surprise is much weaker,
and post-announcement drift much stronger, when a greater number of same-day earnings
announcements are made by other firms. The Industry-unrelated news and large earnings
surprises have a stronger distracting effect. These findings are consistent with the investor
distraction hypothesis. In sum, univariate and multivariate tests provide statistically
significant evidence that high-news days are associated with a lower sensitivity of
announcement abnormal returns to earnings news, a higher sensitivity of post-
announcement abnormal returns to earnings news, and a lower trading volume response
to earnings news.
i) Major studies related to investors overconfidence
During 1987 to 2004, the significant numbers of studies are available related to investor
overconfidence. Some of the major studies have been presented in this part which
includes the overreaction hypothesis, earnings hypothesis, overreaction and decision
76
making, overreaction and trading volume, etc. These studies are classified into three sub-
section based the suitable time frame as below.
i) Major studies related to investors overconfidence before 2000
Table 2.18 organized the major studies and its key findings in 2000 which is presented as
follows:
Table 2.18: Review of major studies on investor overconfidence on stock returns in 2000
Study Major findingsDeBondt and Thaler (1987)
Reconfirmed the overreaction hypothesis i.e. systematic price reversals for stocks that experience extreme long-term gains or losses and, excess returns in January are related to past performance.
Zarowin (1989) The study fails to support the overreaction to earnings hypothesis and concluded the winner-loser effect is primarily a size effect.
Russo and Schoemaker (1992)
The study examines the costs, causes, and remedies for overconfidence and acknowledged that although overconfidence distorts decision making, it can serve a purpose during decision implementation. The overconfidence has remained a hidden flaw in managerial decision making.
Daniel, et.al (1998)
Overconfidence implies negative long-lag autocorrelations, excess volatility, and, when managerial actions are correlated with stock mispricing, public-event-based return predictability.
Odean (1998) Overconfidence increases trading volume and market depth, but decreases the expected utility of overconfident traders.
Camerer and Lovallo (1999)
While analyzing whether optimistic biases could plausibly and predictably influence economic behavior in one particular setting, on undergraduates and MBA graduates. The study concluded that the subjects are simply overconfident; and the inside view which creates that confidence leads them to neglect the quality of their competition.
Hong and Stein (1999)
Each news watcher observes some private information, but failed to extract other news watchers’ information from prices. If information diffuses gradually across the population, prices underreact in the short run but they can only implement simple strategies, their attempts at arbitrage must inevitably lead to overreaction at long horizons.
The study titled Further Evidence On Investor Overreaction and Stock Market
Seasonality, a study made by DeBondt and Thaler (1987) which support the findings of
DeBondt and Thaler (1985) which are systematic price reversals for stocks that
experience extreme long-term gains or losses: Past losers significantly outperform past
winners. The study is based on the major issues regarding the "winner-loser" effect
pertained in the previous work are: first, there is a pronounced seasonality in the "price
correction." Almost all of it occurs in the successive months of January, especially for
the losers. Second, the correction appears to be asymmetric: after the date of portfolio
formation, losers win approximately three times the amount that winners lose. Third, the
characteristics of the firms in the extreme portfolios are not fully described. Finally, the
interpretation of the results as evidence of investor overreaction has been questioned.
Thus, the objectives of the study are: to re-evaluate the overreaction hypothesis and using
the same data set; and, to investigate the hypothesis that the winner-loser effect can be
77
explained by changes in CAPM-betas. The analysis is framed to support the most curious
results - the strong seasonality in the test period returns of winners and losers and a large
portion of the excess returns occurs in January. The study further explores some issues:
first, are there any seasonal patterns in returns during the formation period? Next, within
the extreme portfolios, do systematic price reversals occur throughout the year, or do they
occur only in January? Finally, are the January corrections driven by recent share price
movements or by more long-term factors? Are losing firms particularly small? Are small
firms for the most part losers? Are there any additional excess returns genuinely
attributable to company size when size is measured in a way that is independent of short-
term price movements? Can the study use accounting data to distinguish the overreaction
hypothesis from other explanations of the winner-loser effect? To answer these and other
questions, the study employed the NYSE listed CRSP monthly return data set from 1926
to1982. Calculations of cumulative excess return are made, then ranked and formed 48
portfolios each for winner and loser. The 50 stocks with the highest are assigned to a
winner portfolio while the 50 stocks with the lowest cumulative excess return assigned to
a loser portfolio. Descriptive Statistics, OLS Regression Analysis, Correlation Analysis,
CAPM, Friedman Two-Way Analysis of Variance (chi-square) are used for the analysis.
The principal findings of the study are: excess returns for losers in the test period (and
particularly in January) are negatively related to both long-term and short-term formation
period performance. The winner-loser effect cannot be attributed to changes in risk as
measured by CAPM-betas. The winner-loser effect is not primarily a size effect. The
small firm effect is partly a losing firm effect, but even if the losing firm effect is
removed, there are still excess returns to small firms and the earnings of winning and
losing firms show reversal patterns that are consistent with overreaction. Thus, the
additional evidence that supports the overreaction hypothesis and the seasonal pattern of
returns is also examined where excess returns in January are related to both short-term
and long-term past performance, as well as to the previous year market returns.
Zarowin (1989) on market overreact to corporate earnings information evaluated whether
the stock market overreacts to extreme – good and bad, earnings, by examining firm’s
stock returns over the 36 months subsequent to extreme earnings years. Consistent with
the overreaction hypothesis, stock returns of the poorest earners outperform those of the
best earners. While the poorest earners do outperform the best earners, the poorest earners
are also significantly smaller than the best earners. When poor earners are matched with
good earners of equal size, there is little evidence of differential performance. The basic
78
data analysis strategy of the study is to form portfolios of firms that are characterized by
extreme good vs. bad, current period earnings performance and to compare the
subsequent stock returns of the poorest earners versus the best earners. The CRSP
monthly return file and the Compustat annual industrial file are the sources of the data,
from 1971 to 1981. Stock returns and firm size are computed for the analysis. Firm’s
excess return with market, average excess returns, cumulative average excess return
(CAR), regression analysis and pair analysis are employed. The findings of the study fail
to support the overreaction to earnings hypothesis. The statistically significant differences
between the returns of extreme prior period performers appear to be the result not of
investor overreaction to earnings but of the size effect. The conclusions contrast with
those of DeBondt and Thaler (1987), who maintain, the winner-loser effect is not
primarily a size effect. Thus, the study suggests that size, and not tendency for prior
period losers to outperform prior period winners in the subsequent period.
“To know that we know what we know and that we do not know what we do not know,
this is true knowledge” (Confusius). The good decision marking requires more than
knowledge of facts, concepts, and relationships. It also requires metaknowledge – an
understanding of the limits of our knowledge. Unfortunately, people tend to have a deeply
rooted overconfidence in their beliefs and judgments. Because metaknowledge is not
recognized or rewarded in practice, nor instilled during formal education, overconfidence
has remained a hidden flaw in managerial decision making. Thus, Russo and Schoemaker
(1992) examined the costs, causes, and remedies for overconfidence and acknowledge
that although overconfidence distorts decision making, it can serve a purpose during
decision implementation.
Based on two well-known psychological bases, firstly, investor overconfidence about the
precision of private information and, secondly, bias of self-attribution which causes
asymmetric shifts in investors’ confidence as a function of their investment outcomes.
Daniel, et.al (1998) proposed a theory of securities market under- and overreactions. The
study is guided by the recent years’ of empirical evidences on security returns has
presented a sharp challenge to the traditional view that securities are rationally priced to
reflect all publicly available information. Some of the more pervasive anomalies can be
classified as; even-based return predictability, public-event-data average stock returns of
the same sign as average subsequent long-run abnormal performance; short-term
momentum, positive short-term autocorrelation of stock returns for individual stocks and
the market as a whole, long-run reversal; negative autocorrelation of short-term returns
79
separated by long lags or overreaction; high-volatility of asset prices relative to
fundamentals; short-run post-earnings announcement stock price “drift” in the direction
indicated by the earnings surprise, but abnormal stock price performance in the opposite
direction of long-term earnings changes. The theory is based on investor overconfidence
and variations in confidence arising from biased self-attribution. The study developed the
basic constant confidence model and analyzed the six propositions and showed that the
overconfidence implies negative long-lag autocorrelations, excess volatility, and, when
managerial actions are correlated with stock mispricing, public-event-based return
increase it. When there are many overconfident traders, markets tend to underreact to the
information of rational traders. Markets also underreact to abstract, statistical, and highly
relevant information and overreact to salient, but less relevant information. Like those
who populate them, markets are predictable in their biases. Thus, the study concluded that
overconfidence increases trading volume and market depth, but decreases the expected
utility of overconfident traders.
80
Psychological studies show that most people are overconfident about their own relative
abilities, and unreasonably optimistic about their futures (Weinstein, 1980). Camerer and
Lovallo (1999) analyzed whether optimistic biases could plausibly and predictably
influence economic behavior in one particular setting – entry into competitive games or
markets. The study used the plant-level data from the U.S. Census of Manufacturers
spanning 1963-1982, employed the random and the self-selection sample selection
procedures for 8 different experiments on undergraduates and MBA graduates. The
sample constitutes 118 graduates in total. The study reached to different conclusions
about equilibrium predictions when using the skill-based payoffs instead of random
payoffs – they enter more when betting on their skill, which is not to say that the subjects
behave irrationally – indeed, they forecast the number of competitors quite well, and most
pass tests of expectational rationality. The subjects are simply overconfident; and the
inside view which creates that confidence leads them to neglect the quality of their
competition.
Hong and Stein (1999) developed a model which features the two classes of traders, news
watchers and momentum traders. The study shares the same goal which Barberis et.al
(1998) and Daniel et.al (1998) focused i.e. to construct a plausible model that delivers a
unified account of asset-price continuations and reversals. However, taken different
approach, both earlier studies used representative agent models: Barbaris et.al develop a
regime-switching learning model, where investors wind up oscillating between two states
– one where they think that earnings shocks are excessively transitory and one where they
think that earnings shocks are excessively persistent. Daniel et.al (1998) emphasized the
idea that investors are likely to be overconfident in the precision of their private
information, and that this overconfidence will vary over time as they learn about the
accuracy of their past predictions. While the findings of the study is driven by the
externalities that arisen when heterogeneous traders interact with one another. The
proposed model of the study is judged in terms of three criteria: first, it should rest on
assumptions about investor behavior that are either a priori plausible or consistent with
casual observation; second, it should explain the existing evidence in a parsimonious and
unified way, and finally, it should make a number of further predictions which can be
tested and ultimately validated. With these criteria, the study reached to the conclusion
that each news watcher observes some private information, but failed to extract other
news watchers’ information from prices. If information diffuses gradually across the
population, prices underreact in the short run. The underreaction means that the
81
momentum traders can profit by trend chasing. However, if they can only implement
simple strategies, their attempts at arbitrage must inevitably lead to overreaction at long
horizons.
iii) Review of major studies on investor overconfidence 2000 onwards
The studies on investor overreaction 2000 onwards covered the issues: overconfidence
and assets pricing, overconfidence and trading volume, the survival of pessimism and
optimism in the financial market, gender based overconfidence, etc. Table 2.19 presents
some of the prominent studies and their major findings as follows:
Daniel and Titman (2000) examined why investors are likely to be overconfident and how
this behavioral bias affects investment decisions. The analysis suggested that investor
overconfidence can potentially generate stock return momentum and that this momentum
effect is likely to be the strongest in those stocks whose valuation requires the
interpretation of ambiguous information. Consistent with this, study found that
momentum effects are stronger for growth stocks than value stocks. A portfolio strategy
based on this hypothesis generates strong abnormal returns that do not appear to be
attributable to risk. Although these results violate the traditional efficient markets
hypothesis, they do not necessarily imply that rational but uniformed investors, without
the benefit of hindsight, could have actually achieved the returns. Authors argued that to
examine whether unexploited profit opportunities exist, one must test for what is called
adaptive-efficiency, which is a somewhat weaker form of market efficiency that allows
for the appearance of profit opportunities in historical data, but requires these profit
opportunities to dissipate when they become apparent. In conclusion, the study rejected
the notion of adaptive-efficiency in favor of an alternative theory which suggests that
asset prices are influenced by investor overconfidence.
Table 2.19: Review of major studies on investors overconfidence 2000 onwardsStudy Major findingsDaniel and Titman (2000)
Asset prices are influenced by investor overconfidence.
Barber and Odean (2000)
Overconfidence leads to excessive trading, individual investors who hold common stocks directly pay a tremendous performance penalty for active trading, the overconfidence can explain high trading levels and resulting poor performance of individual investors. Thus, trading is hazardous to your wealth.
Wang (2001) Under-confidence or pessimism cannot survive in financial market, but moderate overconfidence or optimism can survive and even dominate, particularly when the fundamental risk is large.
Barber and Odean (2001)
The study as per the prediction of theoretical models - men are more prone to overconfidence than women, particularly so in male-dominated realms such as finance.
82
Gervais and Odean (2001)
The study developed a multi-period market model, its contribution is explained as: a trader in this model initially does not know this own ability. He infers this ability from his successes and failures. In assessing his ability, the trader takes too much credit for his success. This leads him to become overconfident. A trader’s expected level of overconfidence increases in the early stages of this career. Then, with more experience, he comes to better recognize his own ability.
Jiang et.al. (2004)
High information uncertainty (IU) exacerbates investor overconfidence and limits rational arbitrage.
To shed light on the investment performance of common stocks held directly by
households, Barber and Odean (2000) analyzed a unique data set that consists of position
statements and trading activity for 78000 households at a large discount brokerage firm
over a six year period ending in January 1997. The empirical findings of the study are:
overconfidence leads to excessive trading, individual investors who hold common stocks
directly pay a tremendous performance penalty for active trading, the overconfidence can
explain high trading levels and resulting poor performance of individual investors. Thus,
with the supporting Benjamin Graham’s statement – “the investor’s chief problem and
even his worst enemy is likely to be himself,” the study concluded that trading is
hazardous to your wealth.
The survival of non-rational investors in an evolutionary game model with a population
dynamic for a large economy is analyzed by the study. The dynamic indicated that the
growth rate of wealth accumulation drives the evolutionary process. The study focuses
the analysis on the survival of overconfidence and investor sentiment and found that
under-confidence or pessimism cannot survive, but moderate overconfidence or
optimism can survive and even dominate, particularly when the fundamental risk is large.
Thus, the findings of Wang (2001) provided that new empirical implications for the
survivability of active fund management. The study results lend support to the relevance
of the psychology of investors in studying financial markets.
The study on gender, overconfidence and common stock investment, Barber and Odean
(2001) analyzed the trading behavior of boys and girls. Theoretical models predict that
overconfident investors trade excessively. Psychological research demonstrates that, in
areas such as finance, men are more overconfident than women, this difference in
overconfidence yields two predictions: men will trade more than women, and the
performance of men will be hurt more by excessive trading than the performance of
women. The study tests this prediction by partitioning investors on gender. Thus, theory
predicts that men will trade more excessively than women. Using account data for over
35,000 households from a large discount brokerage, the study analyze the common stock
83
investments of men and women from February 1991 through January 1997. Consistent
with the predictions of the overconfidence models, the study document that men trade 45
percent more than women; the excess trading reduces men’s net returns by 2.65
percentage points a year as opposed to 1.72 percentage points for women. While both
men and women reduce their net returns through trading, men do so by 0.94 percentage
points more a year than do women. The differences in turnover and return performance
are even more pronounced between single men and single women. Single men trade 67
percent more than single women thereby reducing their returns by 1.44 percentage points
per year more than do single women. Thus, the study as per the prediction of theoretical
models - men are more prone to overconfidence than women, particularly so in male-
dominated realms such as finance. Overconfident investors overestimate the precision of
their information and thereby the expected gains of trading. They may even trade when
the true expected net gains are negative, provided the strong support for the behavioral
finance model. Men trade more than women and thereby reduce their returns more so
than do women. Furthermore, these differences are most pronounced between single men
and single women.
It is a common feature of human existence that constantly learns about our own abilities
by observing the consequences of our actions. For most people there is an attribution bias
to the learning: the overestimation of own success. Gervais and Odean (2001) developed
a multi-period market model describing both the process by which traders learn about
their ability and how a bias in this learning can create overconfident traders. A trader in
this model initially does not know this own ability. A trader infers this ability from his
successes and failures. In assessing his ability, the trader takes too much credit for his
success. This leads him to become overconfident. A trader’s expected level of
overconfidence increases in the early stages of this career. Then, with more experience, a
trader comes to better recognize his own ability. Thus, the patterns in trading volume,
expected profits, price volatility, and expected prices resulting from this endogenous
overconfidence are analyzed.
The role of information uncertainty (IU) in predicting cross-sectional stock returns is
examined by Jiang, et.al (2004). IU is defined in terms of value ambiguity or the
precision with which firm value can be estimated by knowledgeable investors at
reasonable cost. The study used different proxies for IU and found that on average, high-
IU firms earn lower future returns i.e. the mean effect and, price and earnings momentum
84
effects are much stronger among high-IU firms i.e. the interaction effect. Thus, the study
concluded that high-IU exacerbates investor overconfidence and limits rational arbitrage.
2.3 Review of major studies in Nepalese context
The history of stock exchange in developed economies has about half a century but it is a
novel practice in most of the developing and transitional economics including Nepal. It is
the general understanding that the maturity of financial market and its participants help to
grow the market substantially. Being the new practice of stock trading in Nepalese stock
market, the quantity of systematic studies seems quite low as well. The major studies on
market information and stock return with its major contributions for the Nepalese Finance
is presented in Table 2.20 as below.
The study examined the relationship of market equity, market-to-book value, price
earnings and dividends with liquidity, profitability, leverage, assets turnover and interest
coverage ratio carried out by Pradhan (1993). The study is based on 55 observations for
the period 1986 to 1990. The sample includes 17 listed companies of NEPSE. Using
linear regression and portfolio analysis to examine the relationship among the variables,
the study revealed that there is positive relationship between market equity and price
earnings ratio, and the negative relationship of market-to-book value with liquidity,
profitability and dividends. Thus, the major finding of the study is the positive
relationship between stock returns and size whereas inverse relation between stock
returns and market-to-book value.
Table 2.20: Review of major studies in Nepalese context
Study Major findings
Pradhan (1993) The positive relation between stock returns and size where as inverse relation between returns and market-to-book value.
Pradhan and Balampaki (2004)
Stock returns is positively related with earning yield and size, where as negatively related to book-to-market ratio and cash flow yield and among the others, book-to-market value was found to be more informative.
Baskota (2007) There is no persistence of volatility in Nepalese stock market and the stock price movements are not explained by the macro-economic variables.
Prasai (2010) The study documented a significant positive relationship between size and stock returns and a significant negative relationship between book to market equity and stock returns.
The study examined the effect of fundamental variables on stock returns and employed
the pooled cross-sectional data of 40 enterprises listed in NEPSE, the only stock exchange
in Nepal. Total 139 observations were collected for the period 1995/96 to 1999/00. The
regression models which explain the stock returns on fundamental variables such as
85
earnings yield, size, book-to-market equity and cash flow yield is employed for the
analysis. The result shows that earnings yield and cash flow yield have significant
positive impact on stock returns and insignificant impact of book-to-market value, size
has a negative impact on stock returns. Among the other findings, the major contribution
of Pradhan and Balampaki (2004) is documented as: the positive relationship of stock
returns with earning yield and size, whereas negative relationship to book-to-market ratio
and cash flow yield and, book-to-market value is found to be more informative.
The impact of trading days, trading volumes, money supply, interest rates, inflation, and
industrial production on the stock returns is analyzed by Baskota (2007). The study is
based on the data collected for the period 1994 to 2006 of NEPSE. There is no persistence
of volatility in Nepalese stock market and the stock prices movements are not explained
by the macroeconomic variables are the findings of the study. Using the event analysis
approach, the study conducted that the political events are not only the factors that
explain the movements in NEPSE.
Prasai (2010) analyzed the fundamental measures and macroeconomic variables and its
influences on stock returns. The sample size for the study is of 48 enterprises listed in
NEPSE whereas total 276 observations were collected. The analysis covered six years
starting from 2000/01 and end with 2006/07. The findings of the study are: a significant
but unexpected positive relationship between size and stock returns; On the other hand,
the study revealed a significant negative relationship between book-to-market equity and
stock returns; while earnings yield and cash flow yield are found to have no predictive
power. The study also examined the individual effect of macroeconomic variables -
interest rate, exchange rate, inflation and money supply and concluded that interest rate
and inflation have the significant explanatory power for stock market movements.
2.4 Concluding Remarks
The stock market movement is one of the most popular areas in finance. The French
mathematician, Louis Jean-Baptiste Alphonse Bachelier is credited with being the first
person to formulate the stochastic model or the random process. Bachelier (1900) the
seminal work is now called Brownian motion. Brownian motion is the presumably
random drifting of particles suspended in a fluid or the mathematical model used to
describe such random movements, which is often called a particle theory. The
mathematical model of Brownian has several real-world applications. An often quoted
example is stock market fluctuations. However, the movements in share prices may arise
86
due to unforeseen events which do not repeat themselves. The Bachelier’s study primarily
focused on; what is the probability that a certain market price be attained before a certain
date? And, the application of probability theory to the stock exchange. The influences for
price movements are innumerable – the past, current and even anticipated events that
often have no obvious connection with its changes might influence the prices. Apart from
the natural variations, some artificial causes might also intervene for price movements.
The stock price movements depend upon infinite number of factors. Thus, it is almost
impossible to predict the market prices accurately using the econometric and the
mathematical modeling. For instance, at the same time, some buyers believe an increase
in the prices whereas sellers trust a decrease. Therefore, it is just an imagination that one
can win with certainty in the stock market. Even if such happened, the combination will
not be persistent because the buyer believes in a probable rise, otherwise he would not
buy, but if he buys, it is because someone sells to him, whereas and the seller obviously
believes in a probable decline. With these explanation, it is logical to state that the
dynamics of the stock price movements is never be an exact science, but it is possible to
study mathematically and with the application of econometric model given that the static
state of the market at a given point of time.
Bachelier (1900) which laid the foundation of stock price predictability and the pioneer
work by Markowitz (1952), the portfolio theory, provided the basis for individual
investors to allocate their resources with due consideration of risk and return tradeoff.
Further, the portfolio theory extended to CAMP which explains the individual stock co-
movements with the overall markets that determine the performance of the stock or the
expected returns which helps to forecast the stock prices. Specially, after the evolution of
CAPM in 1960s, many studies have been carried out to determine the factors affecting the
stock returns. But, the review of major studies suggests that there is lack of consensus on
a single model, methodology and the process of determining the stock returns. For
instance, some evidences shows that stock returns is divided into selectivity and risk
factors whereas others proved that the changes in expected future dividends or expected
future returns leads the stock prices; the firm specific fundamental measures are the
sources of stock returns; the intangible components are the prime causes of stock returns;
the behavioral issues dictates the stock prices; the stock market itself determine its future;
among others, are the major areas of market information and stock returns which have
been continuously contributing for the stock price movements. These empirical evidences
clearly postulated that there are multiple factors that have been supplying variations in
87
stock returns. So that, there is absence of consensus among the existing evidences
regarding the single area of the study, methodology, tools and techniques which clearly
indicates the importance of further studies in this area of Finance.
Stock returns forecasting being the central issues in Finance, numerous studies have tried
to find the most reliable model, tool and technique that explain the majority of variability
of stock returns. To identify such variations, multiple qualitative and quantitative
techniques occupy the major pie of the previous studies. Some studies prioritized the firm
level accounting variables whereas many others documented other variables such as
investor behavior, market behavior, media and political effects, etc. The effects of
individual investor behavior in stock returns have been documented by Lakonishok, et.al
(2011); Loughran and Ritter (1995); French (1980); Brown and Warner (1985); Ritter
(1988), among others use the daily returns whereas Grinblatt and Moskowitz (2004);
Banz (1981); Fama and French (1993, 1996); Chan (2003); La Porta (1996), among
others, employed the monthly returns files for the analysis. Thus, the analysis of the
study would have been extended if the daily and monthly database of the selected
enterprises would be included.
o The study period incorporates the inception of the organized stock exchange operation
in Nepal. But, the study failed to collect the sufficient observations basically before
111
2000 and for the year after mid-July 2010. Former is because of the lack of organized
sources and the later is because of Nepali fiscal year (July 16 th to July 15th). As per the
regulatory provision and the administrative procedures, the listed enterprises are able
to publish its annual financials only after six months or more after completing the
fiscal year. Thus, the study failed to incorporate the latest (2010/011) financials of the
enterprises.
o The number of firm years of the selected enterprises is not similar. Though the study
period is considered as January 13, 1994 to July 15, 2010; the date of listing of the
enterprises, database management, the varying mandatory frameworks, imposed
formatting for regulatory submission for different sector of enterprises is different, etc
cause the variation on observed firm year of the listed enterprises. The variation
ranges from 1 year to 17 years basically as per the age of the firm. Thus, the
survivorship bias or the look-ahead bias as suggested by Fama and French (1996a),
Banz and Breen (1986), and Kothari, et.al (1992) among others is also exists for the
study.
o The primary data is collected from the Kathmandu valley which excluded the opinion
of stock investors residing outside the valley. The investors present during the trading
period at the brokerage floor is considered as the respondents of the study which
bypass the ideas of other investors who perform the stock trading directly from house,
offices and elsewhere. Since, the opinions of the next substantial pie of the
respondents are leaved out for the study so that the results of the primary data analysis
might not the pervasive.
o Similarly, the study used the sample size of 384 stock investors because of the lack of
precise number and list of subjects in the population. The sample size is considered as
364 as suggested by Cochran (1977) at 95 percent level of confidence. Thus, the results
of the study are not free from the limitations of the sample and the sampling procedures.
o The survey was conducted to get the responses of stock investors. Individuals have
different kinds of biases like: Kaniel, et.al (2008) stated that individuals are believed
to have psychological biases; Sum and Wei (2011) suggested the overconfidence
biases; behavioral bias in individuals’ investment choices (Doskeland and Hvide,
2011 and Zhang, 2006 among others), optimistic biases (Camerer and Lovallo, 1999),
Self-attribution biases (Daniel, et.al, 1998), etc. Thus, the study might also be suffered
by similar types of biases which are not tested and considered while the interpretation
of findings. Therefore, the findings of the study would have been better if such types
of biases would have been avoided.
112
o The study analyzed the news effects on stock returns. The published news headings
on a daily newspaper (Kantipur) is assumed a representation sources similar to the
idea suggested by Klibanoff, et.al (1998) and Chan (2003) among others, and does not
consider the other news if any in other news papers. Total news headings for the study
periods are classified into three categories i.e. bad news, good news, and
informational news based on the news heading’s content analysis approach but
excludes the reading of whole article. Thus, the dual interpretable news headings
might mislead the categorization so that such kinds of limitations are essential to
consider while interpreting the results.
o Further, the study ignored the local companies and media effects. Gurun and Butter
(2012) documented that, on average and holding other factors constant, when the
media report news about companies headquartered nearby - that is, local companies -
they use fewer negative words compared to their reports about non-local companies.
Similarly, the study ignored the private information effects on stock returns. Thus, if
the private information effect and the local companies’ effects on media reports would
have been included in the study, the findings might be more glamorous.
o The study incorporates only the political and news effects on stock returns as
intangibles. The other intangibles like: weekend effects (French, 1980); turn-of-the-
year effect (Ritter, 1988); corporate policy changes (Michaely, et.al, 1995);
overconfidence (Odean, 1998; Daniel et.al, 1998; among others); pessimism (Wang,
2001); R&D expenditure and advertising (Chan, et.al, 2001); internet posting and
stock prices (Tumarkin and Whitelaw, 2001); rotation policy of loan officers
(Hertzberg, et.al, 2010); analyst’s recommendations employed by Sun and Wei
(2011); etc are not included. Thus, the results of the study would explain more about
the stock returns if more proxies of intangibles are included in the analysis.
113
Chapter 4
PRESENTATION AND ANALYSIS OF
DATA
This chapter presents the secondary and primary data analysis which deals with the issues
on stock returns and the market information in Nepalese stock market. The chapter is
divided into three sections. The first section is the analysis of secondary database using
the econometric models on the firm specific and market returns series which incorporates
the descriptive statistics and the analysis of stock market returns. The financial news
effect on stock market along with the political leadership effect for the market growth and
development is also shown in this section. The next part shows the presentation of
primary database and the findings of opinion survey, and the third section includes the
concluding remarks on overall data analysis.
4.1 Secondary Data Analysis
In a system approach, the capital market is a component of the whole economic system.
The capital market might be influenced by its own behavior and the other available
information from various sources. In open economic system, the financial market is a
mechanism that fuel for all the economic activities and gradually been influencing by
different kinds of information. The information actually carries some monetary values so
that the valuation of the financial instruments does not remain static for the long period of
time. Being the highly volatile characteristics of the capital market, many opportunities as
well as challenges emerges and disappears if it is not captured at the right time and the
right way. The performance in the capital market in the form of stock price tends to be
useful information for the investment decision makers. But, the magnitude of the
usefulness depends upon the form of the financial market and its growth level. The
financial investors generally use the concept of the investment theories those are
supported by the extensive evidences and, the overall market information can be broadly
classified into fundamental and behavioral information. The accounting growth measures
which is treated as the fundamental information might be the useful market information in
case of relatively static and growing economy. On the other hands, the behavioral issues
114
and the personal characteristics of the stock market participants are also very much
essential to increase the level of confidence and belief towards the sustainable
development of financial market.
The study employed some secondary database to examine the signals of the capital
market movements, different price scaled variables such as: book to market price ratio,
earnings to price ratio, cash flow to price ratio, sales to price ratio, along with the
individual variables and the stock returns series are used as the proxies of market
information because of their relatively static nature. The other information such as:
financial news coverage and political leadership effects are also considered for the
secondary analysis because these variables are not incorporated in the above accounting
growth measures. The news and politics has its own effects for stock returns thus are
treated as separate independent variables and placed them as the other market
information.
Then, the secondary data analysis is presented in a specific order as: the profile analysis,
descriptive statistics, Daniel-Titman regression analysis, news and political effect analysis
for market returns, and, an extended analysis of news and stock returns: the graphical
presentation.
A. Profile Analysis
Table 4.1 incorporates the analysis of 176 enterprises across 14 years starting from 1997
to 2010. Panel A shows the movements of book to market ratio for 825 firm years where
the maximum mean ratio is 1.23 in 1998 and the minimum mean ratio is 0.29 in 2000
followed by 0.35 in 2009, similarly, 1.33 is the highest standard deviation for the year
1998 and the minimum is 0.16 for 2000. Panel B shows the firm year in first row, mean in
second and standard deviation in third row, the figures indicates that the mean book value
of the enterprises gradually decreases from 2000 to 2010 but the movement is volatile
before 2000. The maximum mean book value per share is Rs 293.41 and Rs 145.59 for
the year 1998 and 2007 respectively. The standard deviation on the other hands, indicates
that the highest Rs 215.02 in 1998 and lowest Rs 79.49 in 2009. The Panel C indicates the
average cash dividend of the enterprises and its instablility, as the analysis of 822 firm
years from 1997 to 2010, it ranges between 26.74 percent to 7.16 percent in 2000 and in
2008 respectively where as the standard deviation ranges 49.73 percent to 18.74 for the
year 2010 and 1999 in order. Panel D shows the features of cash flow in million which
115
shows highest in 2009 and lowest (negative) in 1998 for Rs 128.76 million and Rs – 5.10
million
Table 4.1 Profile AnalysisThis table presents the profile analysis of variables: book to market ratio, book value per share, cash dividend, cash flow, cash flow to market price ratio, earnings to price ratio, earnings per share, market equity, market price per share, sales revenue, sales to market price ratio, and stock returns. The measurements employed for the profile analysis are: ratios for Panel A, E (in thousands), F and K (in million); Rs for Panel B, G, and I; millions for Panel D, H and J; and, percentage for Panel C and L. The study period covers from 1997:07 to 2010:07 indicate the firm year, mean and standard deviation for all variables.
respectively and the standard deviation ranges between 35.43 and 888.76. The cash flow
to market price ratio is shown in Panel E which is in general upward sloping, negative in
1998 and highest in 2009, the figures are in thousand. Panel F shows the down ward trend
of earning to price till 2001 and took a upward movement upto 2005 and again decreased
sharply till 2007 and then started to move upwards. In figures, the highest mean ratio is
0.31 early in 1997 and lowest 0.01 in 2007 whereas the standard deviation of earning to
price ratio shows the highest 0.40 in 2007 followed by 0.31 in 2006 and lowest 0.04 in
2001. Similarly, Panel G shows the annual movements of earning per share of 825 firm
year, the general trend indicates the downward movements with some spikes in 2000,
2005 and then showed the upward slope 2007 onwards. The maximum mean earnings per
share is Rs 72.38 in 1997 and minimum Rs 19.94 in 2007 followed by Rs 21.24 in 2006.
The standard deviation on the other hands indicates the high point Rs 63.91 in 2007 and
the low point Rs 28.48 in 1999. Market equity is shown in Panel H which indicates the
upward movement early from the beginning which was started from mean value of
market equity Rs 52.64 million and reached to Rs 588.24 million in 2010 with the same
fashion the standard deviation also started from Rs 59.36 million to Rs 1381.76 million
during the study period. Similarly, the average market price per share is shown in Panel I
which exhibit the U-shape with the highest point of Rs 1055.38 followed by Rs 892.40 in
2000 and 2008 respectively whereas the highest value of standard deviation is Rs 970.09
and lowest is Rs 293.68 for the year 2008 and 1997 respectively. On the other hands,
Panel J shows the yearly features of the sales revenue of 824 firm years which indicates
another U-share during 2000 to 2010 and prior to this period the movement is increasing
till 2000. Panel K similarly indicates the sales to market price ratio, the figures are in
million, the trend line does not shows the smooth movement but potray the ups and
downs where the highest point of mean sales to price is 10.64 million in 2010 and lowest
is 2.30 million in 2008. Finally, the stock returns movements for the period covering 1997
to 2010 is shown in Panel L which constitute 683 firm years with four negative average
return figures. Among the various mean points, the highest is 130 percent in 2000
followed by 103 percent in 2008, after these peak points the stock returns experienced the
sharpe decline. Similarly, the lowest point is negative 32 percent for the year 2010
followed by negative 26 percent in 2002. Standard deviation on the other hands shows the
highest 136 percent in 2008 followed by 78 percent in 2000.
Source: Appendix B
117
The graphical presentation of the figures presented in Table 4.1 are shown in the Figure
4.1 which shows the graph of four price scaled variables and eight other accounting
variables for the period mid-July 1997 to mid-July 2010. The database consist of 176
enterprises and the maximum 825 firm year and the minimum 683 firm year. In aggregate
the movement of majority of the selected variable exhibit the downward movement and
Figure 4.1 Graphical presentation showing the trends of the variables: book to market ratio, book value per share, cash dividend, cash flow, earnings per share, cash flow to market price ratio, earnings to price ratio, market equity, market price per share, sales to market price ratio, sales revenue, and stock returns for the period covering 1997:07 to 2010:07. The figures show the trends of respective variables employed for the study.
Source: Appendix B
118
three out of twelve variables indicate the upward movements namely, market equity
(size), sales to price ratio and the sales revenue.
B. Summary Statistics
Table 4.2 presents the summary statistics of the secondary database which is used to
describe the characteristics of the variables selected for the study. The number of
observation contained maximum of 826 firm years and minimum of 822 firm years
during the study period. Average earnings per share for the whole period is Rs 29.15 with
median value Rs 21.18, the minimum and maximum values are Rs 444.08 (negative) and
Rs 626.00 respectively, the standard deviation is Rs 48.87 and the first and third quartiles
are Rs 11.02 and Rs 36.69 in order. The market price per share in second row shows the
mean value Rs 545.07, the share price ranges between Rs 44 to 6830 during the study
period and the standard deviation is Rs 716.68. Taking the figures of book value per
share, mean is Rs 160.29, median is Rs 138.21, and the maximum and minimum values
are Rs 1005.86 and Rs 364 (negative) in order. The standard deviation shows the
magnitude of variation of the variable which is Rs 97.77 represents the stability of book
value as compare to market price per share. On the other hands, cash dividend represents
the mean value is 11.78 percent with median 1.05 percent and the maximum is 560
percent whereas the standard deviation is 35.23 percent. The figures of sales and cash
flow are shown in sixth and seventh row whereas market equity is in fifth indicates that
the mean value of Rs 287.78 million, minimum value is Rs 8 million and maximum value
is Rs 15000 million, the standard deviation shows Rs 815.04 million. The stock returns is
shown in twelfth row which indicates average returns 5.59 percent with median 2.08
percent, the maximum return is 80.21 percent with 9.54 percent as standard deviation of
the whole study period. The price scaled variables: book to market ratio, earnings to price
ratio, cash flow to price ratio and sales to price
Table 4.2Summary Statistics
This table presents the summary statistics of the variables used for the study. The five point scale with median, standard deviation, number of observations per variables, unit of measurement and the name of the variables are presented in columns and individual variables are shown in rows. The first three variables: earnings per share, market price per share and book value per share are measured in Rs, the cash dividend and stock returns are in percentage terms, the market equity, sales revenue and cash flow are measured in millions in Rs, book-to-market ratio, earnings to price ratio, cash flow to price ratio, and sales to price ratios are measured in times where cash flow to price ratio is in times in thousand and sales to price ratio is in times in millions. All the variables are measured for the period 1997:07 to 2010:07.
This table shows the correlation coefficients of the variables employed for the study which are: earnings per share (EPS), market price per share (MPPS), book value per share (BVPS), sales, cash flow, market equity (ME), stock returns (Rt) and composite share issuance (CSI) variable. The strength of the correlation coefficient is measured at 5 percent level of significance. The Pearson correlation is used for the analysis. The study period ranges from 1997:07 to 2010:07. The figures in parenthesis are p-values.
ratio is placed in eighth to eleventh rows respectively. The mean values are: 0.56, 0.07,
66.21 and 5.59 for the stated price scaled variables respectively where are median values
are: 0.47, 0.06, 5.35 and 2.08, these median values divide the whole series into two equal
parts. Similarly, the standard deviations in number are: 0.53, 0.21, 866.73 and 9.54. The
unit of measurement for cash to price ratio is in thousand and for sales to price ratio is in
million.
Table 4.3 shows the correlation coefficients of the variables are considered for the study.
Among the total correlation coefficients, nine sets of variables which have no significant
correlation and the remaining nineteen pairs have significant positive correlation at 95
percentage confidence interval. Among the significant correlations, the log cash flow and
log market equity has the correlation coefficient 0.61 is the highest value followed by
0.58 for log market equity and log composite share issuance variable whereas the lowest
correlation coefficient is 0.10 which describes the movements of both variables in the
120
same direction. There are total four negative correlations among the variables, out of
them 0.08 (negative) is the highest correlation between log composite share issuance and
the log earnings per share followed by 0.07 (negative) between log book value per share
and log composite share issuance variable. Further, there is negative correlation between
log market equity and log earnings per share and log book value per share.
C. Daniel-Titman Regression Results
I. Book to Market Decomposition
Table 4.4 presents the book-to-market equity decomposition using Fama-MacBeth
regressions. The relationship between the dependent variable and the change in book
value is assumed to have positive whereas market value is negative. The existence of the
stated relationship between the variables proves the information effect on stock price. In
the first regression estimates, the evidence proves that the priori sign is as expected and
significant at 5 percent level. While taking the independent effect of lagged book to
market effect and the changes in book value and changes in market value, the priori sign
is disappeared in model 3 in case of changes in market price. The figures in Panel A
shows that a unit change in BMio leads to 60.80 percent change in Bit/Mit, 0.20 percent
change in △Bi and 0.10 percent (negative) change in △Mi while taking independent
effect the magnitude decreases. The final column indicates the observations retained in
the analysis to generate the Kolmogovor-Smirnov test (p-values) in accepted level. Panel
B shows that the elasticity between the variables i.e. 1 percent change in Log BMi0, Log △Bi and Log △Mi leads to 88.30, 12.00 and 18.60 (negative) percentage changes in
dependent Log Bit/Mit respectively and the explanatory power of the independent
variables is 98 percent as shown in model 4. Thus, the coefficient indicates that the
relationship between the variables is persistence taking the mutual effects of the selected
variables.
Further, the book to market decomposition can be made by replacing the independent
variables and replacing the dependent by firm level stock returns taking 2 to 5 lagged
periods. It is expected to have the same priori sign as presented in Table 4.4. The basic
regression model 3.2 is transformed into three different versions in Panel A, Panel B and
in Panel C which is shown in the Table 4.5 respectively. The first column indicates the
regression models with lagged periods in parenthesis (i.e. 2 to 5) and for each models the
first row shows the coefficients and the subsequent row indicates p-values. Similarly,
seventh column shows the p-values of ANOVA tests, the next column indicates the
121
coefficient of determination, then K-S rest column shows the normality test values (p-
values) and finally, N indicates the number of observations in each model which is ranges
between 403 to 89 observations.
Table 4.4Regression Analysis for Book to Market Decomposition
This table shows the book-to-market decomposition. The dependent variable is log book-to-market equity at time t. The BMio is the lagged book-to-market equity of the firm i at the period 0 to t. △bi is change in book equity at 0 period to t and △mi is change in market price at 0 period to t. R-square is the coefficient of determination, N is the number of observations and the p-values in the Model Sig. column. The study period covers 1997:07 to 2010:07. The p-values are presented in the parenthesis.
α b1 b2 b3 Model Sig R-squareK-S Test of
residual N
Panel A: Log (Bit/Mit) = α + b1 BMi0 + b2△Bi + b3△Mi + ut Model 1 bi -0.640 0.608 0.002 -0.001 0.000 0.95 0.05 437
p (0.000) (0.000) (0.000) (0.000) Model 2 bi -0.484 0.461 0.000 0.84 0.20 279
p (0.000) (0.000) Model 3 bi -0.149 0.000 0.001 0.000 0.17 0.20 279
p (0.000) (0.333) (0.000) Panel B: Log (Bit/Mit) = α + b1 LogBMi0 + b2Log△Bi + b3Log△Mi + ut
Model 4 bi 0.081 0.883 0.120 -0.186 0.000 0.98 0.20 50p (0.000) (0.000) (0.000) (0.000)
Model 5 bi -0.027 0.701 0.000 0.866 0.05 280p (0.002) (0.000)
Source: Appendix B
Table 4.5 presents the regression estimates of the extension of book to market
decomposition. In Panel A, the interpretation is very much similar to the estimates
available in Table 4.4 Panel A but the coefficients are much stronger. For instance, taking
4 lag periods in model 3, the elasticity is 0.914 followed by 0.873 and 0.857 for taking 5
lag and 2 lag periods respectively between log book-to-market and lagged log book-to-
market ratio. Further, 1 unit change in Bt/Bt-i leads to 39.80 percent change, 37.70 percent,
and 33.70 percent changes in log book to market prices by taking 4 lag, 2 lag and 3 lag
periods, in order. Similarly, 1 percent changes in Pt/Pt-i leads to the highest 28.7 (negative)
percent changes in dependent variable in model 1, 19.10 (negative) percent changes in
model 2 and the least effect is 14.50 (negative) percent in model 3.
Table 4.5Regression Analysis for Book to Market Decomposition: An Extension
This table shows the book-to-market decomposition with an extension of firm level stock returns. The dependent variable is log book-to-market equity at time t for Panel A and firm level stock returns from t to t-1 period for Panel B and Panel C. The B t-i/Pt-i
is the lagged book-to-market equity of the firm for the period t-i to t. The Bt/Bt-i is book to lagged book value ratio and Pt/Pt-i is the ratio of price to lagged price at t-i to t period. R-square is the coefficient of determination; N is the number of observations and the p-values in the Model Sig. column, similarly the p-values of K-S test of residuals and dependent variables are presented in the second last column. The study period covers 1997:07 to 2010:07. The p-values are presented in the parenthesis.
α b1 b2 b3 Model Sig R-squareK-S Test of Res/DV (p) N
The R2 values ranges between 96.30 percentages to 82.90 percentages. While replacing
the dependent variable in Panel B by rt-i,t and maintaining no changes in independent
variables, the coefficient indicates the highest 0.23 unit (negative) changes to firm level
stock returns while changing 100 percent in lagged log book-to-market ratio for 3 lagged
periods followed by 0.175 units (negative) in 2 lagged periods. The estimates do not
retain the expected priori signs where it is negative for b1, unstable for b2 and positive for
b3. While explaining the coefficient of b2, it is shown that the highest 0.007 (negative)
unit effects on dependent variable whereas the lowest is 0.001 (negative) unit effects for
firm level stock returns for 3 and 5 lagged periods, in order. Under Panel C, all the
regression coefficients carry the similar interpretation as: a 100 percent change in
independent variables leads to -0.266, -0.094 and 2.006 unit change firm level stock
returns respectively in case of model 9 (taking 2 lag periods). Similarly, in 3 lag periods,
0.281 unit significant changes for lagged log book-to-price, the coefficient is insignificant
for log Bt/Bt-i and 2.299 unit significant changes for log Pt/Pt-i variable. The coefficient of
determination is highest for 5 lagged periods in model 12 followed by model 11. The
analysis is based on 403 observations for model 9 and 95 observations for model 12 and
the K-S test values in Panel B and C are of the dependent variables rather than the
residuals. From the Table 4.5, it is concluded that the firm level stock returns is
negatively affected by the lagged book-to-market ratio and positively by market price to
123
lagged market price ratio but the relationship between returns and the book to lagged
book values is inconclusive.
II. Sales to Price Decomposition
The next regression estimates is the decomposition of sales to price ratio as similar to the
book-to-market decomposition process. The ratio is decomposed into log sales to price
ratio at t-i to t period as base, and the ratio of sales to lag sales and the ratio of price to
lag price for the lag period 2 to 5 years. Table 4.6 shows the estimated parameters of
twelve the regression models employed for different lag periods with the changes and
modification of concerned variables. In each Panel, the estimate starts from the lag period
2 year to 5 years. The dependent variable for Panel B and Panel C is the firm level stock
returns whereas the basic independent variables are the same for all estimates. The
hypothesized priori sign for the parameters b1 and b2 are positive and negative for b3. In
Panel A, all the parameters are statistically significant at 95 percent confidence level. The
coefficient of b1 measure the elasticity between the dependent and log lagged sales to
price ratio where the elasticity is 0.979 for 2 lag, 0.915 for 5 lag, 0.812 for 3 lag and
0.729 for 4 lag periods. But, the coefficients are significantly nil for the variable sales to
lagged sales ratio and the negative relation as per priori for the price to lagged price ratio.
Taking 2 lag period, the coefficient 0.214 (negative) indicates that the movement of price
to lagged price ratio from 0 to 1 leads a 21.40 (negative) percentage change in sales to
price ratio, followed by -0.154 for 3 lag periods and least effect is -0.092 for taking 4 lag
periods. The R-square values ranges from 97 percent to 67.40 percent and the number of
observations in Panel A ranges and 317 to 199 observations. The p-values of K-S test are
the result of the residual analysis. The Panel B shows the relationship between firm level
stock returns and sales to price and its components. From model 5 to model 8, it is shown
that there is minimal and uncertain effect of lagged sales to price effect on stock returns,
no effect of sales to lagged sales ratio and significant
Table 4.6Regression Analysis for Sales to Price Decomposition
This table shows the sales to price decomposition with an extension of firm level stock returns. The dependent variable is log sales to price ratio at time t for Panel A and firm level stock returns from t to t-1 period for Panel B and Panel C. The B t-i/Pt-i is the lagged book-to-market equity of the firm for the period t-i to t. The Bt/Bt-i is book to lagged book value ratio and Pt/Pt-i is the ratio of price to lagged price at t-i to t period. R-square is the coefficient of determination; N is the number of observations and the p-values in the Model Sig. column. The study period covers 1997:07 to 2010:07. The p-values are presented in the parenthesis.
positive effect of price to lagged price ratio for firm level stock returns. More specifically,
0.985 unit changes on firm returns when 1 unit changes in price to lagged price ratio
taking 2 lag periods, 0.964 unit changes, 0.942 unit changes and 0.923 unit changes
taking 5 lag, 3 lag and 4 lag periods respectively and all of them are significant at 5
percent. Further, the Panel C indicates that a 100 percent changes in P t/Pt-i generates 2.533
unit changes in rt-i, t taking 2 lag periods, 2.584 units when taking 4 lag periods, 2.663 unit
changes taking 3 lag periods and 3.62 unit changes in dependent variable while taking 5
lag periods. The coefficient of determination values ranges between 0.998 and 0.977, the
numbers of observations in Panel C are relatively low because of the normality test. The
p-values of K-S test indicate the analysis of residuals in case of Panel A and C and the
analysis of dependent variable in case of Panel B.
In sum, there is consistent negative relationship between sales to price and price to lagged
price ratio and consistent positive relation between firm level stock returns and price to
lagged price ratio whereas inconclusive relation and least effects of lagged sales to price
and sales to lagged sales ratio for stock returns.
III. Cash Flow to Price Decomposition
The decomposition of cash flow to price ratio is shown in Table 4.7 which is divided into
three panels. The normality test is of the dependent variable is shown Panel B and the
residual analysis is shown in Panel A and Panel C. The expected priori sign is proved in
125
Panel A and when the dependent variable changes to firm returns there is consistently
positive sign for price to lagged price ratio. The average coefficient for b1 is 0.585 that
measure the elasticity between cash flow to price and lagged cash flow to price ratio, b 2
constitute 0.024 which describes 1 unit change in cash flow to lagged cash flow ratio
leads to 2.40 percent change in cash flow to price ratio, similarly, the average coefficient
if b3 is -0.163 with the similar interpretation as b2 and the average coefficient of
determination is 51.10 percent. In Panel B, on average 0.993 unit changes in firm returns
because of 1 unit changes in price to lagged price ratio and a100 percent change in lagged
cash flow to price ratio leads to 0.010 unit change in firm returns on an average. An
average R-square is 93 percent and average p-value for K-S test is 0.065 that describe all
the regression models in Panel B are normally distributed. Likewise, the regression
coefficient in Panel C shows the similar meaning as a 100 percent change in lagged cash
flow to price and price to lagged price ratio leads to 0.002 and 1.976 unit changes in firm
level stock return taking lag period 1. Similarly, 0.05, 0.03 and 2.885 unit changes in firm
returns in case of lag period of 2, the figures in regression model 11 are significant at 5
percent. In sum, while taking the independent variable effect, the price to lagged price
ratio has the substantial explanatory power for firm level stock returns during the study
periods.
Table 4.7Regression Analysis for Cash flow to Price Decomposition
This table shows the cash flow to price decomposition in Panel A and the extension of firm level stock returns in Panel B and C. The dependent variable is log cash flow-to-price ratio at time t for Panel A and firm level stock returns from t to t-i period for Panel B and Panel C. The Ct-i/Pt-i is the lagged cash flow-to-price ratio of the firm for the period t-i to t. The Ct/Ct-i is cash flow to lagged cash flow value ratio and Pt/Pt-i is the ratio of price to lagged price at t-i to t period. R-square column indicates the coefficient of determination; N is the number of observations and the p-values in the Model Sig. column. The study period covers 1997:07 to 2010:07. The p-values are presented in the parenthesis.
α b1 b2 b3Model
Sig R-squareK-S Test of Res/DV (p) N
Panel A: log [Ct/Pt] = α + b1 log [Ct-i/Pt-i] + b2 [Ct/Ct-i] + b3 [Pt/Pt-i] + ut Model 0(i=1)
The separation of earnings and price from the variable earnings price ratio is shown in
Table 4.8 which is divided into three Panels. The first section describes the
decomposition of earnings to price ratio into lagged earnings to price ratio and the
independent effect of earnings and price variables. The various studies in the financial
market literatures proved that earnings have
Table 4.8Regression Analysis for Earnings to Price Decomposition
This table shows the earnings to price decomposition with an extension of firm level stock returns. The dependent variable is log earnings to price ratio for the period t to t-i in Panel A and firm level stock returns from t to t-i period for Panel B and Panel C. The Et-i/Pt-i is the lagged earnings-to-price ratio of the firm for the period t-i to t. The Et/Et-i is earnings to lagged earnings ratio and Pt/Pt-i is the ratio of price to lagged price at t-i to t period. R-square is the coefficient of determination; N is the number of observations and the p-values in the Model Sig. column. The K-S test column indicates the test of normality of the series. The study period covers from 1997:07 to 2010:07. The p-values are presented in the parenthesis.
significant effects for price movement in different market context. This table replicates
the similar findings in Nepalese scenario as well. When taking 1, 2 and 3 lag periods,
there is significant effect of lagged earnings to price ratio for firm level stock returns as
shown in Panel B but there is insignificant and negative effect for 4 and 5 lag periods.
Likewise, the earnings to lagged earnings have minimal and insignificant effect for stock
returns but the coefficients of b3 indicates the positive and significant effect on stock
returns. The price to lagged price effect is also strong, consistent and significant in Panel
C but there are unreliable effect of lagged earnings to price ration and earnings to lagged
earnings variables for firm level stock returns as shown in model 11 to model 15. Taking
a look in Panel A, all the regression coefficients are significant at 5 percent level, the
relationship is strong while taking 1 lag periods and as on the increment of the lag periods
to 2, 3, 4 and 5, the magnitude is started to decrease gradually for price to lagged price
ratio, the similar manner for lagged earnings to price ratio except model 5 and the similar
pattern for earnings to lagged earnings ratio. K-S test column indicates the p-values for
normality where Panel A and C is the test of regression residuals and for Panel B, the
values shows the normality test of dependent variable. With these evidences, it is
concluded that for the firm level stock returns there is significant effect of lagged earnings
to price ratio up to three years and the maximum a 100 percent change in lagged earnings
to price ratio generates a 0.032 unit changes in firm level stock returns.
From the Table 4.5 to 4.8, the independent effect of price scaled variables on firm level
stock returns is analyzed. The evidence shows that it is not necessary to increase the stock
returns when sales volume increases because in most of the cases there is negative
relationship between sales to price and stock returns but the relationship is inconsistent
considering the lag periods 2 to 5 years. Another finding is: there is inconsistent
relationship between the firm returns and earnings to price ratio but its strength is more
than the cash flow and sales effect. In numerical form, the maximum effect a 3.2 percent
point change in firm returns because of a 100 percent change in earnings to price variable
128
when lag periods is 2 years and the relationship gradually decreases when lag period is 3
years and fourth and fifth lag periods respectively. When talking about the consistent
independent effects of price scaled variables, it is shown that the positive effect between
firm returns and cash flow to price ratio and the negative relationship of book to market
ratio and the returns. Among the variables, the highest regression coefficients of the
relationship are: 23 percent point changes in firm returns when 2 lag periods and 17.5
percent point changes in firm returns when 3 lag periods because of a 100 point changes
in book to market ratio.
Further, the mutual effect of price scaled variables for the movement of stock returns is
analyzed in Table 4.9. From the table, it shows the disappearance of the negative sign in
most of the cases which is shown in Table 4.5 Panel B and C. Since, the analysis at 4 lag
periods still retains the negative sign but it is only significant at 27.3 percent level. There
is least effect of sales to price and cash flow to price ratio for firm returns but the strong
and significant effect of earnings to price ratio for stock returns is appeared. Taking only
1 to 3 years of lag periods, the numbers of observations are relatively higher as the
maximum is 576 observations and the least is 319 observations. The series is normally
distributed and the coefficients of determination are 28.9 percent, 24.7 percent and 11.4
percent respectively for the lag period 1, 2 and 3 years. The strength of relationship is
significant and consistent for book to market ratio than the earnings to price. When
considering the coefficients of book to price and earning to price ratio for the models 2, 3
and 4, five out of six regression coefficients are significant at two standard deviations. In
contrast, only one regression coefficients in model 5 and model 6 is significant out of ten
coefficients.
Table 4.9Regression Analysis of Firm Returns on Price Scaled Variables
This table shows the analysis of firm returns on price scaled variables. The dependent variable is firm returns from 1 lag to 5 lag years. The independent variables are: book to price ratio, sales to price, cash flow to price and earnings to market price ratio for the period t to t-i. The Bt-i/Pt-i is the lagged book to price ratio for the period t-i to t, S t-i/Pt-i is the lagged sales to price ratio for the period t-i to t, C t-i/Pt-i is the lagged cash flow to price ratio for the period t-i to t, and E t-i/Pt-i is the lagged earnings to price ratio for the period t-i to t. α is constant and the p-values of ANOVA test is Model Sig, R-square is the coefficient of determination; N is the number of observations and K-S test column indicates the test of normality of the series. The study period covers from 1997:07 to 2010:07. The p-values are presented in the parenthesis.
α b1 b2 b3 b4 Model Sig R-squareK-S Test of Res/DV (p)
N
r(t-i,t) = α + b0 [Bt-i/Pt-i] + b1 [St-i/Pt-i] + b2 [Ct-i/Pt-i] + b3 [Et-i/Pt-i] + ut Model 1(i=1)
Thus, it is concluded that for the analysis of firm level stock returns movements only the
three years of historical accounting data are useful to find the signals and to establish the
relationship between the dependent and independent variables.
Table 4.9a presents the regression results of firm level stock returns on book to market,
earnings to price ratio, lagged firm returns and the share issuance variables. The random
terms represent the contribution of undefined variables for the movement of firm returns.
There are altogether eleven models are presented, out of them, the contribution of the
changes in firm’s common stock volume has negligible relation with firm returns because
4 out of 6 coefficients have significantly zero relationship. When looking at the
coefficient of determination, interestingly, when adding one more independent variable in
model 4 (i=3), the r-square value decreased from 25.50 percent in model 4a to 23.50
percent in model 4. The evidence in model 2a and model 5 indicates the extreme
coefficients as 0.30 and -0.30 respectively; coincidently the coefficients are same though
the lag period is different. Against the earlier consistent relationship between the firm
returns and the book to market price ratio in Table 4.5 panel B and C, it is shown the
fluctuating relationship between book to price and returns which is supported by the
irregularly positive and negative signs that appear in Table 4.9i. Best on the proof of
relationship in decomposition section, the book to market ratio is retained and because of
the second strong relationship with firm returns, the earnings to price ratio is also retained
even though its relationship is not consistent. The firm level stock returns and firm returns
are used as interchangeably. The extreme relationship between the firm returns and
earnings to price ratio is 0.499 (positive) and 0.901 (negative) as shown in model 5 and
model 6a, similarly, the book returns’ contribution ranges between 0.703 followed by
0.622 and 0.103 (negative) but only 6 coefficients out of 11 are significant at 5 percent for
book returns.
In majority of the cases, the relationship between past returns and firm returns is negative
which suggest the early winner fail to achieve in later periods and vice-versa. The seven
coefficients are significant at 5 percent among ten for past returns variable. The
interpretation of all the coefficients that appeared in Table 4.9a is: a one unit change in
independent variables (book to price, earnings to price, book returns, past returns and
130
share issuance measure) generates bi unit change in dependent variable. The major
conclusions are: there is fluctuating relation of book to market ratio with firm returns and
there is chances of early losers to achieve in the later periods when taking the analysis of
5 lag periods.
Table 4.9a
Regression Results of Firm Level Stock Returns on Book-to-Market, Earnings to Price, Past Returns and Share Issuance Measures
This table reports the results of a set of regressions of firm level stock returns on lagged book and earnings to price ratios (BP(t-i,t) + b2 EP(t-i,t)), past accounting growth measures – book returns (rB(t-i,t)), past returns (r(t-i,t)) and share issuance measure ι(t-i,t) for the period t-i, to t where ut is the random terms. The study period covers from 1997:07 to 2010:07. The p-values are presented in the parenthesis. R-square is the coefficient of determination; N is the number of observations and the p-values in the Model Sig. column. The K-S test column indicates the test of normality of the series.
The basic regression model 3.2 is being analyzed employing the lagged book to market
ratio as the proxy of expected returns is tangible information and error terms is treated as
the proxy of intangible information. Table 4.10 presents the relationship between firm
returns and book to market ratio with inclusion of book returns. Some key notes are: the
inverse relation of book to market ratio with firm returns once again retained except in
model 1 but there are only half out of total coefficients are significant at 5 percent and the
0.107 (negative) is the highest coefficient followed by 0.071 (negative) in model 3 (i=2)
and model 5 (i=4) respectively. The book returns as an independent variable proved the
positive relationship with returns, 0.164 is the highest value but the model as well as the
coefficient is insignificant followed by 0.078 which is significant (both model and
131
coefficient) in model 6 (i=6) and model 3 (i=2) in order. Four coefficients out of six are
significant at two standard deviations. With these evidences, it is concluded that even
though book returns does not include in the firm returns, it is shown that there is positive
relationship and in some cases the strong relationship with firm level stock
Table 4.10Regression Analysis of Firm Returns on Book-to-Market and Book Returns
This table reports the results of a set of regressions of the firm level stock returns on lagged book to price and book returns (rB(t-i,t)) for the period t-i, to t where ut is the random terms is the proxy of intangible information. The study period covers from 1997:07 to 2010:07. The p-values are presented in the parenthesis. R-square is the coefficient of determination; N is the number of observations and the p-values in the Model Sig. column. The K-S test column indicates the test of normality of the series.
returns. Looking at all the coefficients and the normality of the series, it is shown that the
model 3 (i=2) is the best one where the bi values are relatively strong and significant at 95
percent confidence, the r-square is highest, the model itself is significant and the
observations also are substantial in number (i.e. 285). Thus, it can be said that book to
market ratio is more useful up to 2 lag periods/years which is the more aggressive
findings as compare to the findings of Table 4.9. With the similar procedures, the
usefulness of the other price scaled variables can be calculated.
Table 4.11 gives the detailed analysis of firm returns on price scaled variables with the
extension of fundamental returns for the period July 1997 to July 2010. The regression
coefficients once again proved that there negative relationship between firm returns and
the book to market ratio where as only two coefficients for earnings to price has negative
sign out of eleven regression models for 1 to 5 lag periods. Thus, with the presence of the
fundamental returns measures - the relationship between earnings to price ratio and firm
returns is positive as oppose to decomposition analysis. The values of y1 parameter from
model 1 to model 11 shows the strongest 0.228 (negative) in model 8 (i=4) followed by
model 4 (i=2) where the t-statistics are more than cutoff point at 95 percent confidence
132
level. Contrast to Table 4.9, the predictive power of earnings to price is more than book to
market ratio which can be seen in model 8 (i=4) where the predictive power of earnings to
price is 0.472 unit as compare to 0.228 (negative) unit for book to market ratio. Thus, the
major conclusion of Table 4.11 is: out of fundamentals price scaled variables, only book
to price and earnings to price ratios have strong predictive power and
Table 4.11An Analysis of Firm Returns on Price Scaled Variables with Fundamental Returns Measures
This table presents the regression results of the firm level stock returns on lagged price scaled variables along with the extension of fundamental return measures. From the beginning in the model give below are: book to price, sales to price, cash flow to price and earnings to price ratio and later four variables are: the book returns, sales returns, cash flow returns, earnings returns. All the variables in the model cover 1 lag period to 5 lag period. The study period covers 1997:07 to 2010:07 and the t-statistics are presented in the parenthesis. Coefficient of determination is represented by R-square column; N is the number of observations and the p-values are in Model Sig. column. The K-S test column indicates the test of normality of the series.
the usefulness of the historical data is proved to be the lagged 2 to 4 years where all the
respective regression coefficients are significant at 5 percent risk level. On the other
hands, the sales to price and cash flow to price ratios have no predictive power for firm
level stock returns in Nepalese market as proved by all the coefficients is zero. Similarly,
the fundamental returns measures have least explanatory power for firm level stock
returns. Among them, the book returns has more explanatory power which can be proved
by seeing the coefficients in 5 lag periods (i.e. 0.055 unit) followed by 2 lag periods (i.e.
0.047 unit). The second most predictive power is of earnings returns but there are least
effects of sales returns and cash flow returns respectively. All the presented regression
133
models are significant at 5 percent risk except model 3 (i=1) where F-test p-value is
0.561. The coefficients of determination values are even lesser than it appeared in Table
4.9 where the four independent variables has 28.9 percent explanatory power for the
changes in firm level stock returns but it is only 23.4 percent followed by 22.1 percent for
3 and 2 lag periods in model 6 and model 4 respectively. The data series for all regression
models are normally distributed as shown by K-S test of residuals and the total numbers
of observations are 380 maximum and 158 minimum for model 3 (i=1) and model 10
(i=5) respectively. The holding period stock returns as dependent variable and the proxy
of intangible information as per the regression model 3.9 (Table 4.10) with its extension
to other price scaled variables are treated as independent variables along with the
fundamental to price variables. Share issuance measure is also included as a separate
independent variable in Table 4.12. This table shows the independent effect of intangible
information for the prediction of future stock returns. Two proxies of intangible
information are selected: the first is as per the Table 4.10 and the next proxy is share
issuance measures. The numbers of observations decreases when the lag period increases.
In Panel A, 17/20 coefficients are significant at 5 percent risk and the relationship
between share issuance measures and the holding period returns is nil during whole
analysis (i.e. from lag 1 to lag 5). The direction of the relationship of intangible
information is negative in most cases. Interestingly, fifth year back database is more
useful than the recent database as seen in model 5 (i=5), the coefficient is 0.335
(negative). With this evidence, the historical information which is behind the curtain in
the recent period has significant predictive power than the recent and explicit information
is the finding of this relationship. Once again it is proved that the book value has no
contribution for calculation of holding period returns has to some extent contribution for
stock return movements but the
Table 4.12Regressions Analysis of Holding Period Stock Returns with Intangible Information
This table reports the regression results where dependent variable is holding period firm level stock returns and the independent variables are: fundamentals to price variables, fundamental returns, intangible information and share issuance measure. The B t-i/Pt-i
is the lagged book to price ratio for the period t-i to t, S t-i/Pt-i is the lagged sales to price ratio for the period t-i to t, C t-i/Pt-i is the lagged cash flow to price ratio for the period t-i to t, and Et-i/Pt-i is the lagged earnings to price ratio for the period t-i to t. α is constant. Other independent variables are: book returns (rB(t-i,t)), sales returns (rS(t-i,t)), cash flow returns (rC(t-i,t)), and earnings returns (rE(t-i,t)) respectively for Panel A to Panel D. The share issuance measure (ι (t-i,t)) and the intangible information measure (rI(B) when considering book to price ratio and book returns, rI(S) when considering sales to price ratio and sales returns, rI(C) when considering cash flow to price ratio and cash flow returns, and rI(E) when considering earnings to price ratio and earnings returns (as per Table 4.10 and its extension to calculate the other intangibles) are other independent variables in this table. Model Sig column indicates the p-values of ANOVA test, R-square is the coefficient of determination; N is the number of observations and K-S test column indicates the test of normality of the series. The study period includes 1997:07 to 2010:07. The t-statistics are presented in parenthesis.
Α b0 b1 b2 b3 Model Sig R-squareK-S Test of Residuals(p)
relationship is not consistent during the whole periods. In majority of the cases, the
relationship between book to price and firm level stock return is positive as opposite of
earlier findings but it is only retained till model 3 (or, 3 lag periods). The coefficients in
model 4 and model 5 are stronger than others and significant at 5 percent risk level, thus it
is concluded that initially the book to market ratio contributed positively for stock returns
and the magnitude started to decrease later. The plot of values of b0 parameters gives the
U-shape. Panel B and Panel C is the analysis of sales to price and cash flow to price effect
for holding period returns, the coefficients in most cases is zero and close to zero. Thus,
as earlier, it is concluded that the sales to price and cash flow to price ratio have no
predictive power for firm returns. Similar to Panel A, the intangible information has
negative effect for firm returns in this sub-section. In Panel D, 9/20 coefficients are
insignificant at 95 percent confidence. Likewise Panel A, in some cases, the earning to
price effect is positive and negative then after. Earnings returns have no predictive power.
Thus, the most powerful fundamental return is book returns followed by earnings returns
Source: Appendix B
135
is another findings of the analysis. Once again, the intangible information has negative
relation with firm returns proved by the values of parameter b2. In sum, the major findings
of Table 4.12 is the intangible information in majority of the cases pull down the stock
returns and rarely help to boost the firm level stock returns and when the lag period
increases, the strength of the relationship also inversely increases.
D. News and Political Effect on Stock Returns
News for the study is considered as the newspaper headings of the national daily
“Kantipur” and the classification is made as bad news, good news and informational as
per the content analysis approach. The study period covers 6029 days during 16 years, the
news headings related to the stock market for this period constitutes 1683 headings which
are classified in to 536 bad news, 734 good news and 413 informational news. Table 4.13
shows twelve regression models where the regression model 2 is significant at 9 percent
and the rest models are significant at two standard deviations. The yearly database in
Panel A proves the negative effect of bad news for average market returns similarly, the
informational news contents have also inconsistent and minimal effect for stock returns.
The coefficients of bad news and good news are significant at 5 percent but insignificant
for informational news contents. On the other hands, good news has positive and
significant explanatory power for average market returns during the study periods. The
average coefficient of determination in Panel A is 0.58 where as 0.26 for Panel B and
0.12 for Panel C which indicates that the explanatory power of the yearly database is
higher than monthly and daily. In Panel B, various regression models are formed and
analyzed. Model 4 independently shows the negative relationship of bad news with
average market returns. While adding good news in the model, the value of coefficient
again increases and reaches to -0.010 from -0.008. Similarly, the R-square value also
increases from 21.5 percent to 33.1 percent. Under Panel B, number of months varies
from 134 to 151 and the p-values for K-S test shows the series are normally distributed.
This panel also proves the similar findings as Panel A i.e. the negative effect of bad news,
the positive effect of good news and the inconsistent effect of
Table 4.13News Effect on Average Market Returns
This table presents the regression results between average market returns as dependent variable and the newspaper contents are classified into bad news, good news and information only as independent variables. Panel A shows the yearly effects, Panel B indicates monthly effect and Panel C exhibit the daily news effect on average market returns. The study period covers 1994:07 to 2010:07. In the table, Sig. column indicates the p-values of ANOVA test, R-square is the coefficient of determination; N is the number of observations and K-S test column indicates the test of normality of the series. The t-statistics are presented in parenthesis.
Sources: Data from Appendix C, Appendix D and Appendix F(b)
informational news for the variation of average market returns during the study periods.
The daily database in Panel C also proves the similar conclusion that the daily bad news
adversely affect for the stock returns while daily good news have positive effect for the
market movements and when taking the daily informational news, its contribution seems
marginal and positive. In contrast to the findings in Panel A and Panel B, the daily
informational news effect has significant effect for average market returns even if the
strength is marginal. Thus, the overall conclusion of table 4.13 is that there is negative
effect of bad news contents for the stock market movements where as positive impact of
good news contents and the inconsistent effect of informational news for the market
returns during the whole periods.
The average market index is calculated as the mean of beginning stock market index plus
the closing stock market index and the average market return is calculated similar to the
holding period returns. On the other hands, the mid-July market returns is also calculated
similar to the holding period returns calculations. The analysis in Table 4.14 is classified
into three panels. Panel A shows the analysis of yearly database, Panel B for monthly
database and Panel C describe the relationship between news contents and the market
returns respectively. Regression model 2 among others is insignificant out of twelve
regression models. The average coefficient of determination in Panel A is 0.58 which is
similar to the average stock returns in Panel A of Table 4.13. The findings of this section
137
are: the negative effect of bad news and informational news whereas positive effect of
good news for the stock return movements. In Panel B, the relationships are almost
significant at 5 percent risk level. In most cases, the bad news pushes more than 1 percent
negative changes in stock returns whereas less than 1 percent returns changes and
marginal unit changes (negative) when the newspaper covers informational news. The
number of monthly database in Panel B ranges between 127 months and 149 months and
the average coefficient of determination is 44 percent and the K-S test values indicates
that the series are normally distributed. The coefficients in Panel C indicate that the given
series are normally distributed. All the regression coefficients are significant at 5 percent
risk level. The maximum number of observations is 1689 in model 12. The given
regression models are significant and the coefficient of determination in this panel are
relatively small. The average R-square in this Panel is 3 percent which is very lower than
the similar panel in Table 4.13 (i.e. 12 percent).
Thus, the major conclusion of news effect for stock returns is: there is negative effect of
bad news for stock returns. In most cases one unit of bad news headline leads 0.01 unit
negative change in market returns. The strength of relationship between stock returns and
good news is relatively weaker than bad news but the direction of relationship is
consistently positive i.e. good news leads less than 0.01 unit positive changes in market
returns. The informational news on the other hands has inconsistent and marginal effect
for the stock market movements in Nepalese capital market.
Table 4.14News Effect on Mid-July Market Returns
This table presents the regression results between average market returns as dependent variable and the newspaper contents are classified into bad news, good news and information only as independent variables. Panel A shows the yearly effects, Panel B indicates monthly effect and Panel C exhibit the daily news effect on average market returns. The study period covers 1994:07 to 2010:07. In the table, Sig. column indicates the p-values of ANOVA test, R-square is the coefficient of determination; N is the number of observations and K-S test column indicates the test of normality of the series. The t-statistics are presented in parenthesis.
rm_midjuly = α + b0 bXt + b1 gXt + b2 iXt + ui
Model Constant bXt gXt iXt Sig. R2 K-S NPanel A: Yearly database
Sources: Data from Appendix C, Appendix D and Appendix F(b)
Table 4.15 and Table 4.16 present the relationship between political leadership and its
effect on market returns. Each table has three panels which explain yearly, monthly, and
daily database analysis from 1994:07 to 2010:07. The dependent variable is capital
market returns where yearly database constitute average market returns shows in Table
4.15 (i.e. annual average, monthly average and daily average) and end period (mid-July)
database consist of end of Nepali fiscal year, month end, and the closing daily market
index shows in Table 4.16. For all dummy analysis, the NC led government is treated as
reference group variable.
Table 4.15Political Leadership Effect on Average Market Returns
The table shows the relationship between the average market returns and the political leadership as dummy variable. D1: CPN-UML led government, D2: Other parties led government, D3: UCPN (M) led government where NC led government is treated as the base dummy variable. Panel A shows the yearly database and its effects, Panel B indicates monthly effect and Panel C exhibit the daily political leadership database and its effect. The dependent variable is average market returns. The study period covers 1994:07 to 2010:07. In this table, Sig. column indicates the p-values of ANOVA test, R-square is the coefficient of determination; N is the number of observations and K-S test column indicates the test of normality of the series. The t-statistics are presented in parenthesis.
T (-4.850) (6.863) (-0.478) (4.902) Sources: Appendix E, Appendix F (a) and Appendix F (b)
Table 4.16Political Leadership Effect on Mid-July Market Returns
The table shows the relationship between mid-July market returns and the political leadership as dummy variable. D1: CPN-UML led government, D2: Other parties led government, D3: UCPN (M) led government where NC led government is treated as the base dummy variable. Panel A shows the yearly database and its effects, Panel B indicates monthly effect and Panel C exhibit the daily political leadership database and its effect. The dependent variable is average market returns. The study period covers 1994:07 to 2010:07. In this table, Sig. column indicates the p-values of ANOVA test, R-square is the coefficient of determination; N is the number of observations and K-S test column indicates the test of normality of the series. The t-statistics are presented in parenthesis.
rm_midJul = α + b1D1 + b2D2 + b3D3 + ui
139
Model Constant b1D1 b2D2 b3D3 Sig. R2 K-S NPanel A: Yearly database
Sources: Appendix E, Appendix F (a) and Appendix F (b)
Looking at yearly data analysis in Table 4.15, In Panel A, the CPN-UML led government,
on an average, contributes 0.283 percent negatively in average market returns per year
than those of NC led government while other parties led government contributes 0.104
percent negatively in average market returns per year than those of NC led government.
But, the ANOVA test value (p-value) is much higher than 0.05 thus the fitted model is not
significant even though the series is normally distributed. On the other hand, Panel B and
C give the different evidence that the first, second and third dummies have positive effect
for average stock returns as compare to reference category. Under Panel B, the CPN-
UML, Other parties and UCPN (M) leg government have the positive contribution for
average returns. More specifically, the CPN-UML government, on an average contributes
0.074 percent more than NC government. Similarly, Others parties’ makes on an average
0.027 percent and UCPN (M) makes on an average 0.062 percent more than the NC led
government for the average market returns during 1994:07 to 2010:07 while the daily
database also proves the similar findings. All coefficients in Panel B and C except one
each in both panels are significant at 5 percent significant level and both fitted regression
models are significant at 95 percent confidence. The series are normally distributed and
the numbers of observation are: 148 and 1239 respectively for monthly and daily analysis.
Thus, the major conclusion of this table is: there is lower contribution of the NC led
government for the market growth while CPN-UML and UCPN (M) leadership have on
an average positive contribution for average stock returns.
Further, Table 4.16 exhibit very similar results as Table 4.15 where the fitted regression
model based on year end database is not significant and the monthly and daily data based
regression models are significant at 95 confidence level. The regression coefficients and
its signs are very similar the above results thus, the evidence from Table 4.16 support the
major findings of Table 4.15.
The regression model which contains both the quantitative and qualitative variables is
termed as Analysis of Covariance (ANCOVA) model is shown in Table 4.17. The
140
quantitative variables are: the bad news, good news and the informational news while the
categorical variables are based on political leadership variables as independent variables.
Section A exhibit the average market returns whereas section B shows the end of the
period market returns as dependent variable respectively. Each section constitutes three
ANCOVA models in separate panel. The combination of news count with its
classification and the political leadership with its dummies constitutes 16 years, 153
months and 1245 days of observation for section A and 16 years, 148 months and 1671
days for section B respectively.
Table 4.17A Regression Analysis of Market Returns on News and Political Leadership from 1994:07 – 2010:07
This table presents the regression analysis between market returns (average & mid-July) and the news and political leadership is the dummy variables. D1: CPN-UML led government, D2: Other parties led government, D3: UCPN (M) led government where NC led government is treated as the base dummy variable. Panel A shows the yearly database and its effects, Panel B indicates monthly effect and Panel C exhibit the daily political leadership database and its effect in both sections. Section A is the analysis of average market returns whereas Section B for mid-July market returns. The dependent variables are average market returns and mid-July market returns for Section A and Section B respectively. The bad news, good news, informational news and the political dummies are the independent variables. The study period covers 1994:07 to 2010:07. In this table, Sig. column indicates the p-values of ANOVA test, R-square is the coefficient of determination; N is the number of observations and K-S test column indicates the test of normality of the series. The t-statistics are presented in parenthesis.
Sources: Appendix C, Appendix D, Appendix E, Appendix F(a), and Appendix F (b)
The results suggests that, for every addition of bad news in the selected news paper leads
to mean decrease in average stock returns by about 1.4 percent (-0.014) with t-statics
greater than two standard deviation, for every good news, the average stock returns goes
up by 1.2 percent (0.012) which is significant at 5 percent risk level, and the
141
informational news headings has no significant effect for average stock returns which is
shown in Panel A, Section A of Table 4.17. Meanwhile, taking yearly database, the
political leadership is seen as unbiased or it’s not significant at 95 percent confidence.
When taking the monthly database, there is about equal (0.036 for D1 and 0.035 for D3,
both are significant) positive mean effect as compare to the NC led government. On the
other hands, for average stock returns, the others’ parties led government has similar
effect for average stock returns as compare to the reference NC led government.
Regarding the news effect, similar to Panel A, bad news has negative effect and the good
news has positive effect and both are significant even if the average strength is lower than
that of the coefficients in Panel A. While analyzing the daily effect of news and political
leadership effect on average market returns, the strength of the news (bad news, good
news and informational news) is gradually decreases as shown by -0.014, -0.009 and -
0.003 for yearly, monthly and daily database for bad news, respectively. Similarly, 0.012,
0.007 and 0.002 for good news in Panel A, B and C respectively which describes the
decreasing strength of news when it goes yearly to daily pattern but the t-statistics goes
up meaning the confidence level increases when analysis moves from yearly to daily
database. Thus, the major findings of this section are: the bad news has consistent
negative effect and the good news has consistent positive effect for average stock returns
but there is inconclusive effect of informational news for market returns, the daily news
as well as the leadership effect is more stronger than the monthly and yearly effects, in
general, the CPN-UML and UCPN (M) led government, on an average has positive effect
for the growth of stock market returns.
The section B in the Table 4.17 indicates the news and political leadership effect on end
period market returns. The section is divided into three panels namely, Panel A for the
analysis of yearly database, Panel B for the analysis of monthly database, and Panel C for
the analysis of daily database. The news categories and the political leadership dummies
are the independent variables whereas the dependent variable is mid-July or the end
period market returns. The results shows the consistent negative and positive effect of bad
news and good news for market movements and as oppose to the previous findings in
Section A, the effect of informational news is negative for all panels but it is only
significant at model 3 i.e. daily database. The major differences between the average
market returns and the end period market returns as dependent variable for news effect
analysis are: the yearly end period data have strong effects (-0.014 vs. -0.019, 0.012 vs.
0.015, and -0.001 vs. -0.004) than that of average market returns, the similar results for
142
monthly database but both the strength and the confidence level (-0.009 vs. -0.013, 0.007
vs. 0.010, and 0.000 vs. -0.004 (coefficients), and -6.032 vs. -8.944, 6.245 vs. 8.881, and
0.184 vs. -1.889 (t-statistics)) are higher for end period analysis, and the monthly
database explain more about the end period market returns than that of daily database in
Section A. When talking about the leadership effect for end period market returns, it is
proved that the UCPN (M) led government on an average has more strength to boost the
market growth as compare to NC led government (reference category) and the
coefficients for this variable in Section A and Section B have no differences, the mean
effect of other parties led government has negative effect for end period market returns as
compare to NC leg government under daily database (this result is not significant in
Section A), and, there is no significant differences between NC led government and
UCPN (M) led government for market growth taking the monthly series. But, in sum, the
UCPN (M) led government is placed as more supportive government for market
expansion as compare to NC led government. While looking at the monthly series under
Section B, it can also be interpreted as: irrespective of the absence or presence of news
counts with its categories, the presence of CPN-UML led government is estimated to
increase the end period market returns by an average 3.4 percent; the UCPN (M) led
government is estimated to increase the end period market returns by an average 2.6
percent which is significant at less than 5 percent risk level. But, the coefficients in
Section A (Panel B) are 3.6 percent and 3.5 percent respectively for CPN-UML and
UCPN (M) leg government. Thus, the major conclusions of this section are: the monthly
series have more predictive power than yearly and daily, the bad news, good news and
informational news have consistent negative, positive and negative effect for stock returns
respectively, and the CPN-UML led government is proved to be the stock market friendly
government.
From the above analysis, the major conclusions regarding the news effect are: the bad
news are the market growth barriers and the good news are market growth friendly news
categories, and out of the experienced political leadership in Nepalese context, the CPN-
UML leg government is proved as more constructive government for the market growth
as compare to the NC led government.
E. An extended analysis for the news and stock returns: the graphical presentation.
In Figure 4.2 the pattern of total news along with its categories for the period 1994 to
2010 is presented, moreover, the market returns series are also seen in the upper part of
143
the first graph. Even though the trend line of all the series are upward sloping but the
slope of the market returns
Sources: Appendix C and Appendix D
is steeper than the news counts. Except the period 1999 to 2002, the counts of news
headings follow the market returns patterns. During the whole period, two peaks are seen
where the highest is the corresponding of the fiscal year 2008. The end of the period
market returns patterns shows less smooth movement than that of average market returns
Figure 4.2 The figure below present the graphical presentation of the variables: end of the year market returns, average market returns, total news count and its classification as: bad news count, good news count and informational news counts. Further, the graphs in the second row shows the pattern of average market return (yearly) and the mid-July market returns, and the last table shows the graph of the news categories: bad news, good news and informational news and its total counts. The figures are based on the news headings counts from 1994:07 to 2010:07 and the market returns also based on the same time frame.
144
patterns. Among the news categories, in general, the numbers of good news counts are
greater than bad news and the bad news categories are always higher than the
informational news headings in the selected newspaper for the study. From the graphical
presentation, it is proved that there is linear (positive) relationship between news headings
and the stock market returns, and, good news headings dictates the other categories of
news headings in most of the time during the study period.
Source: Appendix F
The spread of the stock market returns is presented in Figure 4.3. The figure is drawn
based on the daily database of stock market, the initial three years (1994 to 1996) has
limited database due to the unavailability of needful secondary data source representing
the year end index. From 1997 to 2010, the graph is based on the daily database. By the
inspection, the year 2007 experienced the most fluctuation followed by the year 2008.
The graph presents clear two cycles during the study period (i.e. from 1997 to 2002 and
from 2006 to 2010) with clear growth, peak, contraction and trough. Similarly, more or
less the same pattern is followed by the financial news counts.
The insight from the Figure 4.2 and 4.3 poses that when the market parameter tends to
move on the upward basis, the number of financial news count on the national daily also
increases and vis-à-vis but the pattern is not reliable when the number of news count does
not follow the market pattern during 1997 to 2002 but it is from 2006 to 2010. With the
same fashion, when the stock market reached to the peak its spread also poses the highest
Figure 4.3 The figure shows the spread of market returns for the period 1994:07 to 2010:07. Daily market returns is shown in y-axis and time in x-axis. The spikes in the figures indicate the range (minimum to maximum) of the returns for each year.
145
ranges for the year 2000 but the similar inference does not retain for the next cycle. For
the period 2006 to 2010, the market peak is seen in 2008 but the highest spread is seen in
2007. Thus, the inference from the above graphical presentation is that there is no reliable
pattern of the variables measured; there is no guarantee of the linear movement of the
stock market returns and correspondingly the market spread; moreover, the analysis of
news and the market returns also does not clarify whether the news leads to market
returns or market leads to the news counts.
4.2 Primary Data Analysis
The primary data analysis has been classified into two parts, the first part is more focus
towards the demographic characteristics of the respondents and the next section discuss
about the procedure of the factor analysis.
A. Profile of Respondents
Table 4.18 given below exhibit the characteristics of the individual respondents in the
Nepalese capital market in relation to their gender, age, occupation, education, level of
investment and experience of investing. Panel A shows majority of the stock investors are
male and Panel B shows majority are middle age group (i.e. 25 to 40 years) investors. By
profession, the businesspersons involvement is in the first place followed by the service
holders whereas the people who has only the business of investing in the stock market are
placed in the third place considered 22.6 percent of the total respondents. Most of the
investors are well educated as evidence that masters degree & above represent about 44
percent and the investors having bachelor degree are about 37 percent of the total
respondents. From the opinion regarding the level of stock investors, the Panel E
indicates that investors with less than 5 lakh of stock investors constitute 31.1 percent
which indicates that majority of stock participants in are small investors followed the
investors with the investment level 10 to 25 lakh. Panel F presents the investment related
work experience of the survey participants. The experience is classified into five
categories: less than 1 year, 1 to 5, 5 to 10, 10 to 17, and above 17 years. The opinion poll
indicates that majority of the respondents have 1 to 5 years of investing experience i.e.
about
Table 4.18Profile of the respondents based on personal characteristics
This table reports the personal characteristics of stock investors. Panel A indicates the gender, Panel B for age, Panel C for occupation, Panel D for education, Panel E for stock investment (size) and Panel F presents the work experience of the respondents. The numbers of respondents included for the survey are 164.
Variables Demographic Characteristics Number Percentage
146
Panel A: Gender
Female 12 7.3 Male 152 92.7 Total 164 100.0
Panel B: Age of respondents
Below 25 13 7.9 25 to 40 100 61.0 Above 40 51 31.1 Total 164 100.0
Panel C: Occupation
Business 47 28.7 Farmer 2 1.2 Investor 37 22.6 Service 45 27.4 Student 11 6.7 Teacher 9 5.5 Undisclosed 13 7.9 Total 164 100.0
Panel D: Education
Up to high school 10 6.1 Intermediate 13 7.9 Bachelor degree 61 37.2 Master degree & Above 72 43.9 Undisclosed 8 4.9 Total 164 100.0
Panel E: Stock investment (size)
Less than Rs 5 lakh 51 31.1 5 to 10 27 16.5 10 to 25 39 23.8 More than 25 lakh 37 22.6 Undisclosed 10 6.1 Total 164 100.0
Panel F: Experience
Less than 1 year 9 5.5 1 to 5 years 88 53.7 5 to 10 years 43 26.2 10 to 17 years 14 8.5 Above 17 years 5 3.0 Undisclosed 5 3.0 Total 164 100.0
Source: Responses on survey questionnaire in Appendix I
53.7 percent followed by 26.2 percent for 5 to 10 years of capital market related investing
experience whereas the number of new investors who have less than 1 years of stock
investing practice. The number of total respondent are 164 whereas 13, 10, 8 and 5
respondents do not want to disclose their information regarding occupation, level of stock
investment, education level and the work experience respectively.
Table 4.19Investor Education and Personality Profile
This table shows the frequency distribution for investor education and personality profile. Panel A presents the response on the investor’s initial education for stock investment. Panel B indicates the respondents’ interest to participate in investor education program. The Panel C is for the preference towards the investment consultancy and Panel B shows the self-decided personality type of the respondents. Total 164 respondents are included in the following distribution table.
Panel A: How do you learn the basic, how to invest? From….Options Number PercentageFamily Members 19 11.59 Friend Circle 53 32.32 Myself (learning by doing) 49 29.88
147
Education & Training 41 25.00 Undisclosed 2 1.22 Total 164 100.00
Panel B: Do you want to participate in investor education program?
Options N Number Percentage RankTraining program 164 56 34.15 2nd
Discussion forum 164 100 60.98 1st
Interactive talk show 164 31 18.90 3rd
Expo/Exhibition 164 26 15.85 4th
Panel C: Do you want to receive consultancy services?
Options Number PercentageYes 112 68.29 No 41 25.00 Undisclosed 11 6.71 Total 164 100.00
Panel D: What type of investor are you?
Options Number PercentageCautious 26 15.85 Methodical 25 15.24 Spontaneous 12 7.32 Individualist 91 55.49 Undisclosed 10 6.10 Total 164 100.00
Source: Responses on survey questionnaire in Appendix I
The major inferences from Table 4.18 are: the female stock practitioners have just started
to enter into market, the population of middle age people are higher and the education
level of the investors reliably high whereas the volume of investment indicates majority
are small investors having less than 5 years of experience. The evidence of experience as
of less than one year indicates that lately there is less attraction of investors in the market.
Table 4.20Investor Preferences
This table shows the preference of the stock investors towards the market type in Panel A, return preference in Panel B and the factors influencing investment decision in Panel C. The column indicates the options for each survey questions, the frequency of the investor response and the percentage of responses (response # / total respondent * 100). Total numbers of respondents are 164.
Panel A: Which market do you normally prefer?Options Number PercentagePrimary market 17 10.37Secondary market 41 25.00Both 106 64.63Total 164 100.00
Panel B: Returns preference
Options Number PercentageCash dividends 37 22.56Increase in market price 92 56.10Stock dividends 28 17.07Others 1 0.61Undisclosed 6 3.66Total 164 100.00
Panel C: The factors influencing investment decision
Source: Responses on survey questionnaire in Appendix I
Investor education and the personality profile is presented in Table 4.19 where the initial
days of the investors when stepping into the market for stock trading is tried to dig in
Panel A. Based on the survey, 32.32 percent respondents learnt from their friend circle
followed by learning by doing (29.88 percent) where as 25 percent respondents learnt
from their education and trainings. Panel B shows the investors’ preference towards the
investor education program where discussion forum is placed in first position as 60.98
percent preferred this option followed by the training program which is liked by 34.15
percent, the interactive talk show is in third rank and expo/exhibition is the least preferred
among the given options. Majority of the investors preferred to receive the investment
consultancy services as shown in Panel C and finally, Panel D compile the responses of
stock investors regarding their self-decided personality type.
The provided personality types are: cautious- exhibit strong desire for financial security and
is overly careful investor, methodical - Investor who believe on research and rarely form
emotional attachment to the investments, spontaneous - Investor form the portfolio with the
latest hot investment, trade frequently so that the trading cost is too high and individualist -
Investor do the sufficient homework and confident in own abilities where individualist
personality type is placed in the first rank (55.49 percent) followed by cautious
personality types (15.85) and the least preferred once is the spontaneous type. Thus, the
survey finding shows that Nepalese stock investors do the sufficient homework and they
are confident in their own ability while investing in the volatile stock market.
Table 4.21Investor trading behavior
This table shows the investor trading behavior: frequency of monthly trading in Panel A, selection of stock brokers Panel B and the reasons of dealing with more than one broker are presented in Panel C. Total 164 respondents are participated for the poll.
Panel A: How often do you trade securities in a month? Options Number Percentage
0 to 2 40 24.392 to 10 86 52.44Undisclosed 38 23.17Total 164 100.00
Options N Number PercentageTo get more information 164 59 35.98To get share certificate asap on buying 164 14 8.54To collect sum asap on selling 164 43 26.22To develop the public relation 164 36 21.95For convenience 164 14 8.54Do not satisfied with service 164 1 0.61Brokers are less informative 164 1 0.61Others 164 2 1.22
Source: Responses on survey questionnaire in Appendix I
Further, the investor preference based on their judgment on the market preferences, return
preferences, and factors influencing investment decisions are presented in Table 4.20.
Based on the market preference, it is seen that secondary market is the preferred one but
majority of the investors select both types of market mechanism for investment activities.
Panel B on the other hands shows the stock return preference of the stock investors.
About 56 percent respondent preferred the increment in market price followed by cash
dividend 22.56 percent whereas stock dividend is preferred by 17.07 percent of the total
respondents. Media and the friend circle mostly influence the investors’ decision making,
in figure, 50 percent and about 22 percent respectively. Interestingly, only the 9.15
percent of the total respondents do the self analysis while making the investment
decisions. Therefore, the major conclusion of this table is: in the secondary market
majority of the stock investors prefer increase in stock price rather than the fundamental
cash dividend and the media plays crucial role for investment decision making followed
by the friend circle of the stock investors.
Table 4.21 presents the investor trading behavior of stock investors namely, the trading
frequency, selection of brokers and the reasons of selecting more than one brokerage firm
respectively. Panel A shows that majority investors trade 2 to 10 times in a month and
about 24 percent trade below 2 times where the number of investors who do not want this
information is about 23 percent. Panel B shows the selection of brokerage firms for
trading where most of the respondents prefer only to work with one brokerage whereas
32.32 percent investors like to work with two brokers at the same time. The reasons of
dealing with more than one broker are tried to indentify in Panel C where majority of the
respondents stress to get more information (35.98 percent) followed by to collect the
150
money as soon as possible while selling and the third reason based on the survey result is
to develop the public relation. To identify these reasons total 164 respondents and their
responses are analyzed.
Investor trading practices on the various aspects like: trading instruction, execution of
order, time limit, practice of setting time limit, and the time period to receive the share
certificate is presented in Table 4.22. Panel A shows that majority investors i.e. 45.73
percent prefer the limit order or instruct the price range for trading followed by prevailing
market price i.e. 37.20 percent. Panel B indicates majority of trading orders are executed
in the specified price and at the specified quantity. About 63 percent of investors specify
the time limit in order as shown in Panel C while the day order (35.37 percent) followed
by week order (28.66 percent) respectively in numeric form presented in Panel D. The
Panel E gives the clear indication that receiving a share certificate is a very difficult task
in a short time period. Most of the investors i.e. about 84 percent indicates that it takes 20
days and above to get the share certificate after the transaction day.
Table 4.22Investor trading practices
This table reports the investor trading practices. The analysis is classified into five panels. Panel A is about the trading instruction for brokers, Panel B seeks the answer for: how often the orders are executed at specified price and quantity? Panel C is about the time limit in order, Similarly, Panel D is about the practice of setting time limit in order and Panel E is about the time to get the share certificate after transaction day. Total 164 responses are included in this analysis.
Panel A: Trading instructionOptions Number PercentageIn prevailing market price 61 37.20In a limit order (price range) 75 45.73At a fixed price 25 15.24Undisclosed 3 1.83Total 164 100.00
Panel B: How often the orders are executed at specified price and quantity?
Panel C: Do you specify time limit in order?Options Number PercentageYes 103 62.80No 57 34.76Undisclosed 4 2.44Total 164 100.00
Panel D: The practice of setting time limit in order?Options Number PercentageDay order 58 35.37Week order 47 28.66Month order 1 0.61Open/Good-till-cancelled 1 0.61Undisclosed 57 34.76Total 164 100.00
Panel E: How long did it take to get share certificates after transaction day?Options Number Percentage
151
Within 5 days 2 1.225 to 10 days 11 6.7110 to 20 days 12 7.3220 days and above 138 84.15Undisclosed 1 0.61Total 164 100.00
Source: Responses on survey questionnaire in Appendix I
Table 4.23Sources and costs of funds
This table indicates the sources and cost of fund for investors. Panel A seek the answer for How do you manage funds for investment? and Panel B seeks the rate of interest on borrowing in case the investor borrow fund for investment. The numbers of respondents included for the analysis are 164.
Panel A: How do you manage funds for investment?Sources N Frequency Percentage Personal saving 164 103 62.80 Sale of properties 164 11 6.71 Non-interest paying borrowing 164 20 12.20 Interest paying loans 164 61 37.20 Margin loans 164 42 25.61
Panel B: Rate of interest on borrowings
Cost of fund Number PercentageBelow 10 percent 4 2.4410 to 15 30 18.2915 to 20 56 34.15More than 20 percent 6 3.66Undisclosed 68 41.46Total 164 100.00
Source: Responses on survey questionnaire in Appendix I
In Table 4.23, under Panel A, five sources of funds: personal saving, sale of properties,
non-interest paying borrowing, interest paying loans and margin loans. The individual
responses shows that about 63 percent investors use their own personal saving followed
by borrowing interest paying loans (37.20 percent), the third source of fund for
investment is margin loans (25.61 percent) whereas the least preferred source of fund is
sale of properties constitute 6.71 percent. Panel B represent the frequency of responses on
the rate of interest on borrowings. Most of the investors do not want to disclose such
information as shown in table that 41.46 percent of total investors in this categories
whereas 35.15 percent respondent indicates that the rate of interest is 15 to 20 percent
followed by 18.29 percent investors choose the option 10 to 15 percent as the cost of fund
for investing.
The major aim of Table 4.24 is to identify five most preferred information which the
investor usually collect then while making investment decisions. Respondents are
requested to rank their preference as 1st, 2nd, 3rd, 4th and 5th. The collected information are
analyzed and presented in the table above which suggest that dividend is placed in 1 st
152
rank, book value per share in 2nd rank, earnings in 3rd rank, average stock price in 4rd
position and the number of equity is placed in the 5 th position. Moreover, Only 2.4
percent of respondents go beyond the given option as management and 0.6 percent
preferred economic growth as major factors influencing investment decisions making.
Thus, the major findings of Table 4.24 is the most important five information that
investors usually consider prior investing are dividends, book value per share, earnings,
average stock prices, and the number of equities.
Table 4.24The preferred information prior investing
This table presents the list of preferred information prior investing. The first column gives the selected list of information generally considers prior to investment decision and the rest column indicates the investors ranking in number and in percentage. The other information if investors considered for making investment are placed in undisclosed row. The numbers of respondents included in this analysis are 164.
Source: Responses on survey questionnaire in Appendix I
Table 4.25 presents the investor risk perception on the various issues like: risk preference,
management of risk, optimal number of enterprises for diversification, and the time for
revision of portfolio. Based on the responses of the stock investors in Panel A, the
changes in fundamentals or the firm specific variables is the most important variable for
generating risk in investors mindset. The change in monetary policies is placed in second
position. Similarly, the changes in capital market policies through the end of SEBON is
placed in the third row and changes in macro-economic factors in forth. Finally, news and
the media coverage are placed in the least risk generating factor whereas about 14 percent
respondents are not interested to disclose this information.
Panel B shows the opinion results on the management of the risk on investment where the
option – investing in different sectors and different companies is the most preferred as 77
percent respondents like this option followed by 17.7 percent for investing in same sector
153
and different companies. The least preferred option is investing in only a single company.
In Panel B, the optimal number of companies for well diversification is 5 to 10 companies
(44 percent) followed by 10 to 20 companies (30.5 percent). Panel C in the next part
shows investors generally change their portfolio once in 3 months.
Table 4.25Investor's risk perception
This table presents the investor’s risk perception due to the changes in the micro and macro economic factors along with the news & media effect and due to the role of regulatory authority in Panel A. Panel B collect the opinion of investors for managing risk on investment. Panel C shows the number of optimum companies for better diversification and the Panel D indicates the tentative time taken for the revision of portfolio. In total 164 responses are collected and analyzed for this analysis.
Change in capital market policies (SEBON) 20 12.2 5th 27 16.5 3rd 30 18.3 2nd 31
18.9
1st 37 22.6 2nd
Undisclosed17 10.4 26 15.9 26 15.9 26
15.9
26 15.9
Total 164 100 164 100 164 100 164 100 164 100
Panel B: Managing risk on investment
OptionsNum
. %
Investing in same sector, different companies 29 17.7
Investing in different sectors, different companies 126 76.8
Investing in only a single company 1 0.6
Spontaneous buying 2 1.2
Averaging price 2 1.2
Undisclosed 4 2.4
Total 164 100
Panel C: For diversification, what is the optimal number of enterprises?
Below 5 companies 12 7.3
5 to 10 companies 72 43.9
10 to 20 companies 50 30.5
20 and above companies 24 14.6
Undisclosed 6 3.7
Total 164 100
Panel D: Revision of portfolio
Once in 3 months 67 40.9
Once in 6 months 44 26.8
Once in a year 27 16.5
Do not revise 21 12.8
Undisclosed 5 3.0
Total 164 100
Source: Responses on survey questionnaire in Appendix I
Table 4.26Investor's perception and awareness level
The table shows the investor’s perception and the investor’s level of awareness. Panel A indicates the investor’s
154
perception on MF, CDS, CRA and PMS (where, MF = mutual fund, CDS = central depository system, CRA = credit rating agency, and PMS = portfolio management service). The scale for this panel is: not important, less important, neutral, important and most important (in order). Similarly, Panel B measures the awareness level using the scales as: not aware, less aware, neutral, aware and highly aware respectively. Total numbers of respondents for this analysis are 164.
Panel A: Investor's perception
OptionsMF CDS CRA PMS
Number % Number % Number % Number %Not important 9 5.5 9 5.5 8 4.9 10 6.1Less important 7 4.3 8 4.9 11 6.7 8 4.9Neutral 14 8.5 10 6.1 19 11.6 21 12.8Important 58 35.4 51 31.1 53 32.3 59 36.0Most important 61 37.2 70 42.7 49 29.9 47 28.7Undisclosed 15 9.1 16 9.8 24 14.6 19 11.6Total 164 100 164 100 164 100 164 100
Source: Responses on survey questionnaire in Appendix I
In Panel A, the MF and CDS are considered as the most important mechanism for the
market development where 37.2 percent respondents and 42.7 percent respondents select
the option ‘most important’. Similarly, 32.3 percent for CRA and 36 percent for 36
percent of the total respondents select the option ‘important’ for market growth and
development. Panel B on the other hands, measure the investor awareness level on the
stated MF, CDS, CRA and PMS where most of the investors preferred to select the option
‘aware’ on all of the stated mechanism followed by ‘highly aware’ option but 20 percent
of the total investors state that they are not aware regarding CRA. Thus, the major finding
of this table is that even though investors perceive MF, CDS, CRA and PMS are most
important mechanism for market growth and development but they are not highly aware
on any of them so that it indicates the need of investor education on these specific areas.
Table 4.27Investor Reactions on capital market issues
This table shows the trading strategies in different situations. There are 10 statements/situations are given with the buy/sell/don’t trade option. The last column indicates the number and percentage of the undisclosed respondents. In all rows, total numbers of respondents are 164.
Situations NBuy Sell Don’t trade Undisclosed
Number
% Number % Number % Number %
a) Before the AGM of the enterprise (at the end of the fiscal year) 164 78 47.6 28 17.1 41
25.0
17 10.4
155
b) When the AGM is likely to announce the cash dividend 164 101 61.6 23 14.0 26
15.9
14 8.5
c) When the AGM is likely to announce the right shares 164 49 29.9 55 33.5 40
24.4
20 12.2
d) When the AGM is likely to announce the bonus shares 164 123 75.0 16 9.8 10 6.1 15 9.1
e) When the quarterly reports of the enterprise show increase in profit volume
164 107 65.2 11 6.7 3018.
316 9.8
f) When the enterprise launch the new product/service in the market 164 44 26.8 10 6.1 88
53.7
22 13.4
g) When the enterprise issue bonds in the market
164 17 10.4 30 18.3 9557.
922 13.4
h) When the media frequently cover the positive news of certain enterprise
164 93 56.7 16 9.8 3521.
320 12.2
i) When the media coverage highlight the merger news of the enterprises
164 21 12.8 29 17.7 8551.
829 17.7
j) In case if financial institution, when NRB going to declare crisis ridden institution
164 2 1.2 128 78.0 2012.
214 8.5
Source: Responses on survey questionnaire in Appendix I
The investor reactions regarding the trading strategies on the various market situations are
presented in Table 4.27. The total numbers of respondents are included for this analysis is
164 where most of the investors (i.e. 75 percent) want to buy the stocks when the
upcoming AGM of the enterprise is likely to announce the bonus shares, followed by
about 65 percent investors want to buy the stocks when the quarterly report of the
enterprises show the increase in profit volume. Similarly, when the AGM is likely to
announce the cash dividend about 62 percent investors want to buy those stocks and very
few (only 1.2 percent) investors want to buy stock in case of financial institution, when
NRB is going to declare crisis ridden institution. Under the selling strategy, majority (78
percent) of stock investors want to sell their stock in case of financial institution, when
NRB is going to declare crisis ridden institution, followed by 55 percent of total
respondents want to sell their stocks when the upcoming AGM is likely to announce the
right shares, similarly, 30 percent of total respondents want to sell their stock when the
enterprise issue bonus share in the market. In the next column, the do not trade strategies
and its frequency is presented in the table where majority (58 percent) do not want to
trade when the enterprise issue bonds in the market, followed by 54 percent investors
don’t want to trade when the enterprise launch the new product/service in the market.
Table 4.28 exhibit the investor judgment on the various issues and the evidences of the
previous studies. In all rows, the maximum number of respondents is 160 (in Panel A)
and the minimum is 155 (in Panel B). In Panel A, investors agree on the only one issue
156
i.e. ‘news events lead some investors to react quickly’ with the mean value 1.170 and for
rest four issues, most of the investors raise their opinion towards the disagreement. While
looking at mean figure, the highest in the disagreement section is 1.875 followed by 1.792
and 1.731 and 1.658 respectively. With these opinions, it is concluded that investors are
poses disagreement on the statement ‘investing in IPO is more risky than investing in
secondary market’, ‘seasonal offerings do not maximize the shareholders’ wealth’, ‘the
most frequent trading is harmful for investors; wealth’, and ‘if reliable private
information is available, it would be better to invest in single security’ and they tend to
agree on the statement ‘news events lead some investors to react quickly.’ Under Panel B,
the investor judgment on the evidences of the previous studies is presented where the
likert scale is designed into 4 points. The values allocated for the scale are as: strongly
agree (1), agree (2), disagree (3) and strongly disagree (4). Out of the total respondents
only 96.3 percent investors respond for statements b and d. The mean value if greater than
2 indicates the responses move towards the disagreement section whereas when mean
value is less than 2 indicates the responses move towards the agreement section. The
highest mean value is 2.523 shows that most of the respondents express their
disagreement on ‘high information uncertainty enhance the investor’s overconfidence’
followed by the mean value 2.380 for ‘media effect, market noise, seasonal effect, etc
strongly enhance the investor’s overconfidence’ where as the mean value for ‘stock
market exhibit higher returns following good news and lower on bad news’ is 1.904
indicates the investors are agreed on the essence of the evidence. Thus, based on the
views of the stock investors in Panel B, it’s concluded that investors are agreed on the
evidence ‘stock market exhibit higher returns following good news and lower returns
following bad news’ and disagreed on ‘investors under-react to publicly available
information and overreact to perceived private information’; ‘investors respond
mistakenly in initial phase of the information disclosure’; ‘the media effect, market noise,
seasonal effect, etc strongly influence men investor but not for women’ and ‘high
information uncertainty enhance the investor’s overconfidence’ where the magnitude of
agreement and disagreement are in the order as the statements written above.
Table 4.28 Investor Judgment on various issues and evidences
This table presents the investor judgment on the various issues and evidences of the studies. Panel A incorporates the investor judgment on the various issues with clearly indicating the number of respondents in second column. The mean value is presented in the next column and final column indicates the rate of respondents on each row out of total 164 respondents. Panel B further extend the scale into 4 points and its statements are the evidences of the various studies in the area of capital market in the international arena.
Panel A: Investor judgment on the various issues
157
Statements N MeanAgree Disagree I don't know Total
%Num. % Num. % Num. %
a) Your judgment on "investing in IPO is more risky than investing in Secondary market" (Loughran and Ritter, 1995)
160 1.875 22 13.4 136 82.9 2 1.2 97.6
b) Your judgment on "seasonal offerings do not maximize the shareholders' wealth"
160 1.731 48 29.3 107 65.2 5 3.0 97.6
c) Your judgment on "if reliable private info., it would be better to invest in single security"
158 1.658 60 36.6 92 56.1 6 3.7 96.3
d) Your judgment on "the most frequent trading is harmful for investors' wealth"
159 1.792 42 25.6 108 65.9 9 5.5 97.0
e) Your judgment on "news events lead some investors to react quickly" (Klibanoff, et.al, 1998)
159 1.170 139 84.8 13 7.9 7 4.3 97.0
Panel B: Investor judgment on the various evidences
Statements N Mean
Strongly agree Agree Disagree Strongly disagree Total
%Num. % Num. % Num. % Num. %
a) Your response on "stock market exhibit higher returns following good news and lower on bad news" (Zhang, 2006)
157 1.904 52 31.71 74 45.12 25 15.24 6 3.66 95.7
b) Your response on "media effect, market noise, seasonal effect, etc strongly influence men investor but not for women"(Biais et.al, 2005)
158 2.380 33 20.12 42 25.61 73 44.51 10 6.10 96.3
c) Your response on "high information uncertainty enhance the investor's overconfidence" (Jiang et.al, 2004)
d) Your response on "investor under-react to public info. and overreact to perceived private information" (Chan, 2003)
158 2.259 31 18.90 66 40.24 50 30.49 11 6.71 96.3
e) Your response on "investors respond mistakenly in initial phase of the information disclosure" (Ikenberry et.al, 1995)
156 2.340 26 15.85 59 35.98 63 38.41 8 4.88 95.1
Source: Responses on survey questionnaire in Appendix I
Apart from structured questionnaire asked for the respondents, some open questionnaires
are also provided to collect the views of respondents without any restrictions so that some
investors raised the issues which they considered the most important factors for market
growth and development. For instance, most of the investors who respond on the open
question stated that political stability is one of the crucial factor for enhancing investor
confidence and market growth followed by the implementation of central depository
system and they perceive that Nepalese stock market is influenced by the big investors.
Similarly, the other issues are: investor awareness, need of proper financial analysis,
investor confidence, the need of modernization, the policy implementation, etc among
others.
B. Factor Analysis
158
This part of the study is directed towards to identify the key components that influence
the stock prices in the market. The factors those affect the stock prices need to be
analyzed while investing or to succeed in the market, the investors needs to look at them
prior to making investment decisions. The factor analysis is based on the survey database
with initially 16 different variables but to maintain the factor analysis procedure, 2
variables are omitted so that the remaining 14 variables are included in the procedure
after the anti-image correlation matrix.
Since, the basic procedure is the preliminary screening of the responses through the
correlation analysis. The factor analysis is designed in such a way that the included
variables should have to optimum level of relationship among the other variables. Table
4.29 shows that there are 120 correlation coefficients where 42 coefficients are significant
at 5 percent risk level. Because of relatively the small sample size, the correlation
coefficients in the table do not meet the perfect requirements of the factor analysis so that
it is essential to increase the number of respondents for the study. By skipping the
correlation matrix analysis, the study jump to next step the measure of sampling adequacy
to overcome the existing limitations.
Table 4.29Correlation Matrix and p-values
This table presents the correlation matrix and the p-values in the succeeding table below. The variable are defined as: analyzing financial statements is not important (X1), analyzing the rate of price changes is an important step for trading securities (X2), brokers usually alter my investment decisions (X3), graphs, lines & charts are useful for stock trading (X4), I always evaluate the company profile & track records of management while investing (X5), I believe that success in stock market depends upon luck (X6), I use the average prices to determine the current prices (X7), I use dividend payment records while buying and selling stocks (X8), the price move in a direction provides insight about future prices (X9), It is important to look at debt and equity structure before investing (X10), News/media largely influence my investment decisions (X11), Political instability is not the major cause of stock market downturn (X12), analyzing high-low prices is important while buying and selling stocks (X13), Macro-economic indicators and monetary policy dictates my investment decisions (X14), my friends recommend/help me to decide most of my investment alternatives (X15), and, I do not perform the proper financial analysis myself while investing (X16). 5 point Likert scale technique is used to collect the responses where 1 indicates ‘strongly agree’ and 5 indicates ‘strongly disagree’. The Pearson’s correlation coefficients are tested at 5 percent level. Total 164 respondents are included in the opinion collection. The coefficients in the table below are p-values. There are 164 responses included for the analysis.
Source: Responses on survey questionnaire in Appendix I
In Table 4.30 the diagonal of the MSA table represent the MSA values which are greater
than 0.50, the benchmark value as per Kaiser’s recommendation. The MSA result shows
that the sample is adequate for performing the factor analysis. Thus, there is way out to
proceed to the next step.
Table 4.30Anti-image Correlation Matrix
This table reports the Anti-image correlation matrix. The variable are defined as: analyzing financial statements is not important (X1), brokers usually alter my investment decisions (X3), graphs, lines & charts are useful for stock trading (X4), I always evaluate the company profile & track records of management while investing (X5), I believe that success in stock market depends upon luck (X6), I use the average prices to determine the current prices (X7), I use dividend payment records while buying and selling stocks (X8), the price move in a direction provides insight about future prices (X9), It is important to look at debt and equity structure before investing (X10), News/media largely influence my investment decisions (X11), analyzing high-low prices is important while buying and selling stocks (X13), Macro-economic indicators and monetary policy dictates my investment decisions (X14), my friends recommend/help me to decide most of my investment alternatives (X15), and, I do not perform the proper financial analysis myself while investing (X16).
Source: Responses on survey questionnaire in Appendix I
In the table 4.31, the KMO’s MSA test shows that the measure of correlation pattern in the sample is 0.654 considered as good for the further analysis. Bartlett’s test of Sphericity which is the test of null hypothesis of no correlation among the variables under consideration but it is rejected at 95 per confidence level so that the fundamental requirement for the factor analysis is fulfilled.
Table 4.31KMO and Bartlett's Test
The table presents the Kaiser-Meyer-Olkin Measure of Sampling Adequacy coefficient and Bartlett's Test of Sphericity with approximate chi-square value, degree of freedom and the p-value. The test is performed to confirm the sampling adequacy. The analysis is based on the responses collected from 164 respondents.
Kaiser-Meyer-Olkin Measure of Sampling Adequacy. 0.654
Bartlett's Test of Sphericity Approx. Chi-Square 264.530
Df 91
Sig. 0.00
Source: Responses on survey questionnaire in Appendix I
Table 4.32 presents the major aspects of the factor analysis procedure. The initial 14
variables with its factor loadings, extraction, specific variance, and the initial and rotated
Eigenvalues are shown in the table. Each factor loading indicates the relationship between
the corresponding -
Table 4.32An initial solution for factor analysis
This table shows the initial solution of the factor analysis, communalities coefficients of the variables for the data reduction procedure. The first column includes the variables selected for the study. The extraction column indicates the power of variation explained by the corresponding variables. The specific variance is column indicates the variation causes by the other components beyond the corresponding stated variable. The extraction method: Principal Component Analysis (PCA) is used. The components coefficients indicate the correlation between the individual variable and the selected components. The last four rows shows the initial Eigenvalues and the percentage of variance explained by the components. The rotated Eigenvalues as well as percentage of variance explained is also presented in last two rows.
Source: Responses on survey questionnaire in Appendix I
variable and its component, extraction is the power of variance explanation, the specific
variance is the contribution of other components except the corresponding variable, and
the Eigenvalues are the basis for finding the reliable components in the factor analysis
procedures. For example, the variable: X5: “I always evaluate the company profile &
track records of management” explain 47 percent of the total variance through its
aggregate contribution on five components. On the other hand, the specific variance is the
proportion that is not explained by the stated variables. In other words, 53 percent of the
total variation is not covered by X5 variable. The similar interpretation is applicable for
remaining variables and its coefficients in different columns like: X9: the prices move in a
direction provides insight about the future prices extract about 81 percent of total variance
whereas only about 19 percent of the total variation explain by the remaining components
i.e. except stated five components. Initial Eigenvalues in the forth last row indicates the
sum of square of each factor loadings in each component. For instance, the first
component has the capacity to explain the total variance of about 18 percent whereas the
fifth component has 7.42 percent. When the five components combine together that
constitute the total variance explanation power of about 58 percent. When, the factor
analysis proceed to the varimax rotated solution the total variance explaining power
remains constant i.e. 58.396 percent but the individual component’s capacity tend to
changes or come to quite uniform way. The individual Eigenvalues of the rotated
solutions and its percentage of variance explanation become changed. The negative factor
loadings show the negative relationship between the variables and the components.
162
Figure 4.4
The Scree Plot and the determination of number of components greater than 1 Eigenvalues
Source: Responses on survey questionnaire in Appendix I
The justification of the selection of five principal components can be made through two
ways. First one is the straight criterion; under this criterion the number of principal
components is equal to the number of Eigenvalues greater than 1. The next criterion is the
scree plot; the number of components is equal to the number of Eigenvalues greater than
first scree. The Figure 4.4 exhibits the scree plot of the Eigenvalues but the determination
of the reliable components does not clearly shown. Generally, even though the existence
of the different criteria, in most cases both methods determine the same number of
163
components. For this study, five components are determined prioritizing the first criteria
for this study.
Table 4.33 presents the rotated solution with cross loadings in Panel A and the final
rotated solution in Panel B after omitting the three cross loading variables. The factor
loadings are suppressed which are smaller than 0.40 and sorted them as per the retained
components. Finally, only 8 components are retained with three factors. The
independence test is presented in the later part of this section.
The rotated solution in Panel A has 14 variables with 5 components. The results suggest
that there are 3 cross loadings in the table so that the necessary treatment is need to be
done. The best solution for the cross-loading is considered as the omitting such variables
from the remaining factor analysis procedure. The cross-loading variables are omitted
following one by one procedure.
Table 4.33Rotated solution for factor analysis
The table below presents the rotated solution for factor analysis having 5 components and 14 retained variables in Panel A and 3 components and 8 retained variables in Panel B. The PCA method is used for extraction and Varimax with Kaiser Normalization is used as rotation method. The factor loadings are suppressed below 0.40 and ranked in ascending order.
X14 The macro-economic indicators and monetary policy dictates my investment decisions 0.709 X5 I always evaluate the company profile & track records of management while investing 0.625 X10 It is important to look at debt and equity structure before investing 0.620 X6 I believe that success in stock market depends upon luck -0.525 0.423 X4 Graphs, lines & charts of stock market indicators are useful for stock trading 0.488 0.475 X7 I use the average prices (6 months, 1 yr, 2 yrs, etc) to determine the current prices 0.792 X8 I use dividend payment records while buying and selling stocks 0.731 X3 Brokers usually alter my investment decisions 0.776 X11 Media coverage largely influence my investment decisions 0.661 X13 The analysis of high and low prices is important while buying and selling stocks 0.482 X16 I always perform the proper financial analysis myself while investing 0.723 X15 My friends recommend/help me to decide most of my investment alternatives 0.696 X1 Analyzing financial statements is important for trading securities -0.462 0.405X9 The prices move in a direction (increasing/decreasing) provides insight about future price 0.880
Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization.Rotation converged in 7 iterations.
Panel B: Rotated Component Matrix
StatementsComponents
1 2 3
X3 Brokers usually alter my investment decisions 0.768 X11 Media coverage largely influence my investment decisions 0.652 X15 My friends recommend/help me to decide most of my investment alternatives 0.587 X8 I use dividend payment records while buying and selling stocks 0.839 X7 I use the average prices (6 months, 1 yr, 2 yrs, etc) to determine the current prices 0.788 X10 It is important to look at debt and equity structure before investing 0.820 X5 I always evaluate the company profile & track records of management while investing 0.677 X9 The prices move in a direction (increasing/decreasing) provides insight about future price 0.457
164
Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization.Rotation converged in 5 iterations.
Source: Responses on survey questionnaire in Appendix I
After omitting the cross-loading variable from the factor procedure, the Panel B shows
the final rotated solution where 8 variables are retained with 3 components/factors. The
first factor constitutes three variables (X3, X11 and X15) namely, “brokers usually alter
my investment decisions”, “media coverage largely influences my investment decisions”,
and “my friends help/recommend me to decide most of my investment alternatives”.
Similarly, the second factor constitutes two variables: “I use dividend payment records
while buying and selling stocks”, and “I use the average prices to determine the current
prices.” Finally, the third factor includes three variables (X10, X5, and X9) namely, “it is
important to look at debt and equity structure before investing”, “I always evaluate the
company profile & track records of management while investing”, and “the prices move
in a direction provides insight about future prices”. Thus, the most important step in
factor analysis is to name these factors incorporating the features of all the included
variables so that based on these criteria, the study determined the name of these factors
as; the external factor (brokers, media & friends) for the first factor, self-knowledge
(using dividend records & average prices), and the firm specific factor (debt and equity,
company and management profile & price movement). Therefore, the factor analysis
concluded that the external factor, self-knowledge and firm specific factor are the most
important factors among others that directly influence the investment decision making
procedure.
Table 4.34Correlation matrix of the retained variables and factors identified: A verification
This table reports the correlation matrix of the retained variables until the final procedure in Panel A and final outcomes of the factor analysis in Panel B. The Pearson’s correlation coefficients show the degree of relationship between components and the p-values are presented in the row corresponding p below each correlation coefficient.Panel A: Correlation Matrix of variables under consideration: For verification
Panel B: Correlation Matrix of factors identified: For verification