7 - Chapter 4

CHAPTER 4
BENJAMIN GRAHAM’S STOCK SELECTION CRITERIA
4.1 INTRODUCTION
The major objective of investors is to maximize their expected return and
reducing its related risks (Hejazi and Oskouei, 2007). In order to maximize the returns of
investors, Graham and Dodd (1934) developed a few sound principles for analyzing a
company‟s fundamentals and its future scenario. They revolutionized the investment
theory by introducing the concept of security analysis, fundamental analysis and value
investing theory. Value investing utilizes a traditional fundamental analysis approach in
selecting stocks for investment portfolios (Ahmed, 2008). The basic premise of value
investing is to invest in stocks trading below their intrinsic value. The difference between
the stock‟s intrinsic value and its market value is called as margin of safety. According to
Graham, an investor must invest in the stocks which trade at significant discount to their
intrinsic value so that the margin of safety can protect him in the event of a huge fall in
price. Thus, Graham and Dodd (1934) through their seminal work entitled “Security
Analysis” enlightened the investor community with the concept of margin of safety as the
core principle of investing in stock market. Graham (1949) in his another seminal work
entitled “The Intelligent Investor” suggested rules for stock selection for defensive as
well as enterprising investor. In his last years, he distilled his six decades of
understanding into ten criteria that would help the intelligent investor to pick value stocks
from the chaff of the market. These rules/ principles given by Benjamin Graham as listed
in Rea (1977), Oppenheimer (1984), Klerck and Maritz (1997), Phalon (2001) are as
under:
1. An earnings to price yield should be at least twice the triple-A bond yield. The
earnings yield is the reciprocal of the price to earnings ratio.
2. A price to earnings ratio should be lesser than 40% of the highest price-earnings
ratio the stock had over the past five years.
3. A dividend yield should be of at least two-thirds of the triple-A bond yield.
4. A stock price should be lesser than two-thirds of tangible book value per share.
73
Benjamin Graham’s Stock Selection Criteria
5. A stock price should be lesser than two-thirds of net current asset value (current
assets less total debt).
6. Total debt should be lesser than the book value.
7. Current ratio (current assets divided by current liabilities) should be greater than
two.
8. Total debt should be lesser than twice the “net current asset value”
9. Earnings should have shown a compounded annual growth of at least 7% in last
ten years.
10. Stability of growth in earnings i.e. there should be no more than two declines of
five percent or more in year-end earnings over the most recent ten years.
The studies examining the relevance of stock selection criteria of Graham have
found considerable premium yielded by these principles (see section 2.1 in chapter 2).
However, Most of these studies relate to U.S. and other mature markets. The Indian stock
market, on the other hand has become comparable to other mature markets due to various
financial sector reforms initiated since early 1990s by the Government of India. The
positive fundamentals combined with fast mounting markets have made India a striking
end for foreign institutional investors. It therefore becomes imperative to examine the
relevance of a value yielding investment strategy in Indian stock market. The present
study attempts to examine the relevance of stock selection criteria of Benjamin Graham
in Indian stock market for the period 1996-2010.
It is attempted here to explore the profitability of the criteria by screening stocks
on the basis of said principles and then the significance of excess returns, if any, yielded
by such stocks is examined. The excess of raw returns over market index (Sensex)
returns, that is, market adjusted returns and the abnormal returns determined through
capital asset pricing model have been used for the purpose of analysis. The sample size
for the study ranges from 66 companies for the year ended 1996 to 252 companies for the
year ended 2010 (as described in section 3.2 of chapter 3). Given below is the discussion
as well as empirical testing of each and every principle of stock selection of Benjamin
Graham.
74
4.2 TEN PRINCIPLES OF STOCK SELECTION- A THEORETICAL

FRAMEWORK AND EMPIRICAL ANALYSIS
4.2.1 P1. Earnings Yield: The earnings to price yield should be at least twice the
AAA bond yield
The earnings yield is the inverse of the most commonly followed valuation metric
price to earnings (P/E) ratio. It is calculated by dividing the most recent 12-month period
earnings per share by the current market price per share. The price to earnings ratio is
also called as „multiple‟ for the reason that it denotes how much an investor is ready to
pay for 1 of its earnings. A stock is trading at a P/E of 5 means that the investor is
ready to pay 5 for 1 of its earnings. Thus, a high P/E ratio is the indication of high
earnings growth expected out of stock in future. Therefore, this valuation metric
establishes the relationship between the actual recent earnings based performance of the
company with its future market performance.
Graham and Dodd (1934) questioned the ability of the firms to sustain same
growth in earnings in future, so they hypothesized that firms who have and are currently
experiencing high (low) earnings growth are unlikely to able to sustain it to the extent
expected by the market e.g. a high price to earnings multiple is indicative of the market‟s
expectation of high future earnings growth. When this earnings growth reverts towards
industry/ economy mean, then this will result in the revision of earnings‟ expectations, a
fall in firm‟s price to earnings multiple and so a downward correction in its stock price
(Bird and Gerlach, 2003). Therefore, it is wise to concentrate on portfolio of stocks
whose prices are depressed while depicting excellent value at the same time. These
securities must portray excellent value at present in order to create a buffer against future
market volatility. Thus, regardless of market volatility, the value of such a portfolio
remains intact in short term. Most importantly, over the long term there is strong potential
for this portfolio to increase (Ahmed, 2008). Graham recommended that the yield on the
earnings should be at least twice the AAA bond yield (Graham, 1949). This requirement
meant that the qualifying stock‟s P/E could be no more than 1/2r, where r is the AAA
corporate bond rate, measured in decimals. If AAA corporate bond rate were 10%, the
P/E could be no more than 5 [1/ (2*0.1)]. This relationship had to be true in order to
75
compensate the risk that that the earnings might fall (Au, 2004). Therefore, the higher
earnings yield of stocks designate them to be undervalued relative to bonds.
Seeing the great depression of 1929, Graham strongly believed that the stocks are
riskier than the bonds due to the fact that at the time of liquidation of the company, the
bondholders are first in the queue to get back the money and the shareholders are at last
(Anderson, 2012). So in order to ensure the wide margin of safety to investors, Graham
recommended that the stocks should have at least double the yield on bonds to protect the
investors against the loss or unease in the event of some future decline in net earnings
(Graham, 1949).
It can be seen that the earnings yield has been compared with AAA rated bonds.
AAA rating is the highest rating assigned to an instrument and such assets are deemed
least likely to default (Marshall, 2009). Instruments with the AAA rating (by CRISIL) are
considered to have the highest degree of safety regarding timely servicing of financial
obligations and an issuer of such security has very strong capacity to meet up its
financial commitments1. However, during the subprime mortgage crisis of 2007 in the
U.S, the bonds or securities which were rated as AAA were downgraded to CCC by
rating agencies on account of lack of the issuer‟s ability to meet its financial commitment
(Olofsson, 2008). Therefore, instead of taking AAA bond yield, the yield on government
security has been taken as they carry least risk of default and, hence, are called risk-free
gilt-edged instruments2. Rate of returns on 91-days Treasury Bills has been used as a proxy for
risk free return (Assoe and Sy, 2004; Tripathi, 2009; Campbell et al., 2009).
For decades, the investment analysts have used earnings yield and its inverse
(price to earnings ratio) to determine whether the stock is undervalued or overvalued. The
securities that trade at high price to earnings multiple is characterized as overvalued or
growth stocks and the stocks that sell at low price to earnings multiple are classified as
undervalued or value stocks. Some analysts use absolute measures to categorize stocks as
value stocks, for instance, the authors have divided the stocks on the basis of their P/E
ratio into terciles, quartiles, quintiles, deciles etc. to classify stocks as value stocks and
growth stocks. Based on P/E ratio, the stocks in the highest tercile, quartile, quintile or
decile are called as growth stocks and in the lowest segment are called as value stocks.
1
http://www.crisil.com/ratings/credit-rating-scale.html
2
http://www.rbi.org.in/Scripts/FAQView.aspx?Id=79
76
Studies have explored the potential of earnings yield (or P/E ratio) in yielding
value premium and explaining the variation in the equity returns all over the globe
(Nicholson, 1960; Basu, 1977; Reinganum, 1981; Basu, 1983; Goodman and Peavey,
1986; Jaffe et al., 1989; Chan et al., 1991; Claessens et al., 1995; Brouwer et al., 1996;
Mukherji et al., 1997; Bauman and Miller, 1997; Bauman et al., 1998; Fama and French,
1998; Anderson and Brooks, 2006; Moller, 2009; Truong, 2009; Al-Mwalla et al., 2010).
They found that the stocks selected on the basis of low P/E ratio tend to outperform the
stocks with high P/E ratio on the basis of risk adjusted returns.
However, several studies (Ball, 1978; Fama and French, 1992; Mukherji et al.,
1997; Campbell and Shiller, 2001; Kyriazis and Diacogiannis, 2007; Hejari and Oskuei,
2007; Ibrahim and Nor, 2011) have found the lesser explanatory power of earnings yield
variable as compared to other valuation variables in explaining the overall return.
Moreover the literal principle of earnings yield given by Benjamin Graham has limited
exploration. Only a few of the studies like Oppenheimer (1984), Klerck & Maritz (1997)
have examined the role of this principle in providing returns greater than that of market
index returns. Hence, an attempt is made in the present study to screen the stocks, each
year, from 1996 to 2010, satisfying the said principle. Then, the market adjusted returns
(excess of raw returns over market index returns) of such stocks are calculated and the
following hypothesis is tested.
H01 : The mean/median market adjusted return of the stocks having
earnings to price yield at least twice the 91 days Treasury bill yield is
equal to zero.
In order to test the above hypothesis, the stocks having earnings to price ratio of at
least twice the 91 days Treasury bill rate (taken as proxy for AAA bond yield in present
study) are screened on 30th June every year, from 1996 to 2010. The stocks so arrived
have been held for the period of 12 months as well as 24 months (as described in section
3.4 of chapter 3). The Table 4.1 reports the results of one sample t-test and one sample
Wilcoxon signed rank test3 employed to examine the significance of returns in case of 12
months, 24 months holding period.
3
Wilcoxon signed rank test is the nonparametric equivalent of one-sample t-test with the null
hypothesis that the median value of the market adjusted returns of the stocks in the sample is equal
to zero (Hussein, 2005).
77
Table 4.1: Results of the Significance of Market Adjusted Returns of Stocks

having Earnings Yield at least Twice the AAA Bond Yield
12 Months Holding Period 24 Months Holding Period
No. of
Year Mean Std. t- p- Mean Std. t- p-
stocks
(Annual) Dev. value value (Annualized) Dev. value value
-56.6528 -35.9683
1996 4 32.80548 - .041** 35.47269 - .144
(16.402) (17.73635)
1997 -10.8569 -5.4154
16 31.66028 -1.372 .190 28.74470 -.754 .463
(7.9150) (7.18617)
1998 2.8141 -5.5279
27 39.79658 .367 .716 29.00992 -.990 .331
(7.6588) (5.58296)
32.0802 6.9592
1999 18 95.53962 1.425 .172 41.72090 .708 .489
(22.5189) (9.83371
8.9804 38.0899
2000 32 52.04169 .976 .337 27.45214 7.849 .000***
(9.1997) (4.85290)
72.1239 36.2971
2001 53 56.93013 9.223 .000*** 25.05987 10.545 .000***
(7.81996) (3.44224)
27.1575 15.6825
2002 30 37.70336 3.945 .000*** 23.25311 3.694 .001***
(6.8836) (4.24542)
14.0812 12.4759
2003 51 39.12184 2.570 .013** 19.69012 4.525 .000***
(5.47815) (2.75717)
34.6917 5.6487
2004 68 58.06299 4.927 .000*** 26.41710 1.763 .082*
(7.04117) (3.20354)
-13.4622 -8.4204
2005 46 50.22380 -1.818 .076* 25.83708 -2.210 .032**
(7.40510) (3.80947)
.0472 9.5775
2006 23 37.04970 .006 .995 27.60627 1.664 .110
(7.7254) (5.75630)
11.0801 8.9806
2007 21 68.80814 .738 .469 33.00812 1.247 .227
(15.015) (7.20296)
27.0269 24.3016
2008 30 66.49993 2.226 .034** 25.54813 5.210 .000***
(12.1411) (4.66443)
41.4789 9.6900
2009 117 37.82360 11.862 .000*** 24.90116 4.209 .000***
(3.4967) (2.30211)
-10.5522 -8.6279
2010 68 38.69873 -2.249 .028** 27.77644 -2.561 .013**
(4.6929) (3.36839)
Across
21.2181 9.3983
the 604 55.22851 9.442 .000*** 30.06865 7.682 .000***
(2.24722) (1.22348)
period
Note:
1. *, **, *** denotes p-values significant at 10, 5 and 1 percent level respectively
2. Standard error of mean has been reported in parenthesis.
3. Italicized values represent the p-values of Wilcoxon signed rank test
From Table 4.1 we notice that, the number of stocks meeting the criterion of
earnings to price at least twice the 91 days Treasury bill yield ranges from 4 to 117 across
the period of 15 years. The stocks so arrived have been providing positive market
adjusted return in 11 out of 15 year period in case of 12 months holding period. However,
the positive market adjusted returns have not been significant in all 11 years due to larger
78
standard deviation of returns from the mean. The positive mean market adjusted reruns
have been significant at 1% level of significance for the year; 2001, 2002, 2004 and 2009.
For the year; 2003 and 2008, the positive mean market adjusted return has been
significant at 5% level of significance. However, the criterion provides significantly
negative market adjusted returns only in 3 years i.e. 1996, 2005 and 2010. Further, across
the period of 15 years, the stocks selected on the basis of earnings to price ratio of at least
twice the risk free yield, provides mean market adjusted return of 21.21%, which is
significant at 1% level of significance. Thus, the stocks selected on the basis of this
principle enable an investor to acquire positive market adjusted returns in 11 years out of
15 years period and significantly positive returns in 6 years when the stocks have been
held for the period of 12 months each year.
Also, it is evident from Table 4.1 that when we extend the holding period of the
stocks from 12 months to 24 months, the criterion provides us positive mean market
adjusted returns in 10 years out of 15 year period. Out of those 10 years, the market
adjusted returns have been significantly positive in 2000, 2001, 2002, 2003, 2004, 2008
and 2009. Thus in 7 years, the market adjusted returns have been significantly positive
when the holding period has been extended from 12 months to 24 months period.
However in years; 1996, 2005, 2010 the returns remain significantly negative even after
extending the holding period to 24 months. Overall, across the period of 15 years, the
stock selection based on first rule of Graham helps an investor to reap the mean market
adjusted annualized rate of return of 9.39%, which is significant at 1% level of
significance. Thus, an investor can acquire significantly higher returns than the market by
relying on the principle of earnings yield being twice the risk free yield. Hence, the
applicability of this principle cannot be ignored in the present day scenario in Indian
stock market.
4.2.1 P2: Price Shrinkage: A price to earnings ratio should be lesser than 40 per
cent of the highest price-earnings ratio the stock had over the past five years
A price to earnings ratio should be lesser than 40 per cent of the highest price-
earnings ratio the stock had over the past five years. Graham believed that the stock
79
selection should not be based solely on the current price to earnings performance of the
company. Also, the value of past record of the firm‟s performance as an indicator of
future earnings should not be ignored. Moreover, the accounting figures taken for the
fairly long period of time (5 years in context of this rule) tend to absorb and even out the
distorting influences of the business cycle (Graham, 1934). So, a stock having P/E ratio
lesser than 40% of its highest P/E ratio over the last five years is considered undervalued.
In other words, the current price-earnings ratio of a stock has to be down to at least 60%
of the highest price-earnings ratio of the stock over the past five years.
Since P/E multiple differs across sectors, with stocks in some sectors consistently
trading at lower P/E ratios than stocks in other sectors, one cannot judge the value of a
stock by comparing its P/E ratio to the average P/E ratio of stocks in the sector in which
the firm operates. Thus, a technology stock that trades at 15 times earnings may be
considered cheap because the average P/E ratio for technology stocks is 22, whereas an
electric utility that trades at 10 times earnings can be viewed as expensive because the
average P/E ratio for utilities is only 7 (Damodaran, 2004). Thus, the price-earnings ratio
of the stock has been compared with its own 5 year highest P/E only because P/E could
be different across different industries since it is affected by changes in debt levels and
tax rates (Greenblatt, 2006). Hence, the following hypothesis is tested:
H02 : The mean/ median market adjusted return of the stock selected on the
basis of principle of current price to earnings ratio (P/E) of the stocks
being lesser than the 40% of the highest P/E that the stock had over
the past 5 years is equal to zero.
In order to test the above hypothesis, the stocks meeting the second principle have
been screened every year from 1996 to 2010. Then, the market adjusted returns of the
stocks meeting the above principle have been examined using one sample t-test statistic.
Further, the returns have been analyzed taking 12 months as well as 24 months as the
holding period of the stocks. Table 4.2 reports the results of one sample t-test employed
to examine the significance of returns in case of 12 months, 24 months holding period.
80

having Current P/E Lesser than 40% of the Highest P/E the Stocks
had over Past 5 Years
12 months holding period 24 months holding period
No. of
stocks
-45.7539 -40.6434
1996 34 43.64442 -6.113 .000*** 40.75680 -5.815 .000***
(7.48496) (6.98973)
-11.7509 -10.0845
1997 30 66.23686 -.972 .339 36.01142 -1.534 .136
(12.09314) (6.57476)
6.8737 -4.1120
1998 48 39.70704 1.199 .236 28.35682 -1.005 .320
(5.73122) (4.09295)
14.6506 2.8407
1999 43 74.88141 1.283 .207 35.42256 .526 .602
(11.41931) (5.40189)
3.4291 32.9339
2000 39 49.25231 .435 .666 25.78582 7.976 .000***
(7.88668) (4.12904)
71.1648 35.6441
2001 49 57.31298 8.692 .000*** 25.42323 9.814 .000***
(8.18757) (3.63189)
16.3335 3.3038
2002 24 36.16123 2.213 .037** 30.93038 .523 .606
(7.38138) (6.31364)
1.1057 6.6032
2003 35 44.11741 .148 .883 23.90341 1.634 .111
(7.45720) (4.04041)
44.0015 9.7384
2004 36 64.83409 4.072 .000*** 27.75733 2.105 .043**
(10.80568) (4.62622)
-16.6064 -13.9403
2005 17 45.04200 -1.520 .148 28.75562 -1.999 .063*
(10.92429) (6.97426)
4.2455 .3248
2006 11 40.37243 - .859 26.07343 - .657
(12.17274) (7.86143)
-11.1678 14.5663
2007 24 59.58628 -.918 .368 22.80634 3.129 .005***
(12.16300) (4.65532)
27.9985 23.5423
2008 46 56.54682 3.358 .002*** 23.21658 6.877 .000***
(8.33737) (3.42310)
38.8515 7.4220
2009 48 35.24445 7.637 .000*** 23.61329 2.178 .034**
(5.08710) (3.40828)
-6.8737 -10.1252
2010 47 34.51399 -1.365 .179 25.85414 -2.685 .010**
(5.03438) (3.77121)
Across
13.2080 5.6720
the 531 58.56729 5.197 .000*** 34.07012 3.836 .000
(2.54160) (1.47852)
period
Note:
2. Standard error of mean has been reported in parenthesis
81
As apparent from Table 4.2, the stocks meeting the criterion of having P/E ratio
lesser than the 40% of the highest P/E ratio over the past 5 years, ranges from 11 to 49
across the period of 15 years. Further, the stocks so arrived in the portfolio provide
positive mean market adjusted returns in 10 years out of 15 years period in case of 12
months holding period. Further, the returns have been significant in 2001, 2004, 2008
and 2009 at 1% level of significance and at 5 % level of significance in case of year
2002. However, only in year 1996, the criterion provides significantly negative market
adjusted returns. Nonetheless, across the period of 15 years, the portfolio provides
significantly positive market adjusted return of 13.2% in case of 12 months holding
period.
Also, it can be seen from Table 4.2 that, when the holding period of the stocks
selected on the basis of second principle of Graham, is extended to 24 months, the stocks
provide positive mean market adjusted returns in 1999, 2000, 2001, 2002, 2003, 2004,
2006, 2007, 2008, 2009. Thus, out of 15 year period, in 10 years, the mean market
adjusted return has been positive and in year; 2000, 2001, 2004, 2007, 2008 and 2009, the
returns have been significantly greater than zero. Also, the returns have been significantly
lesser than the market in 1996, 2005 and 2010. On the whole, the principle provides mean
market adjusted annualized rate of return of 5.67% (significant at 1% level of
significance) across the period of 15 years in case of 24 months holding period.
Thus, across the period of 15 years stocks selected on the basis of this principle
have yielded significantly positive market adjusted returns in both of the holding period.
The principle of P/E ratio of the stocks to be lesser than 40% of the highest P/E the stock
had over past 5 years, however untouched by academicians, still makes its place in Indian
market on the basis of positive market adjusted returns yielded to investors in maximum
number of years.
4.2.1 P3: Dividend Yield: A dividend yield should be of at least two-third of the
AAA bond yield
When a corporation earns the profit or surplus, it can either re-invest its earnings
in the business or distribute among its shareholders. The portion of the earnings
distributed to the shareholders is called as dividend. A flourishing company is one which
can pay dividends on a regular basis and generally increase the rate as the time lapses. An
82
inductive study would undeniably show that the earning power of companies does not in
general increase in proportion with the increase in the retained surplus, thereby implying
that the common stocks which pay liberal dividends sell higher than other stocks with the
same earnings power but paying out only a small part thereof. Thus, given two companies
in the same general position and with the same earning power, the one paying the larger
dividend will always sell at the higher price (Graham and Dodd, 1934).
Graham recommended that dividend yield should be at least 2/3 of the bond yield.
Since, the bond is considered as safer and steady source of earnings, the investor would
observably be skeptic regarding investment in stocks. Thus, in order to invest in a stock,
its dividend yield should be at least 2/3rd of bond yield (Graham, 1949). This prerequisite
ensured that the security has to be competitive with bond as an earnings generating
device. Since bonds are intrinsically safer than stocks, one would expect stocks to
generate superior return over bonds. Graham however believed that if a company assures
the payment of dividend and also has the prospects of moderate future growth, then a
dividend yield that started at two third of the bond yield would eventually exceed the
fixed income stream, leading to capital gains as well (Au, 2004).
Moreover, Stocks with high and apparent sustainable dividend yields that are
competitive with high quality bond yields may are more resistant to a decline in price
than lower-yielding securities because the stock is in effect “yield supported”4.
Also, worn out by the poor market of recent years, many investors have reached
the end that retained earnings bear only a unsubstantiated relationship to subsequent
capital gains and thus believe that the high dividend yielding stocks offer on a risk
adjusted basis, greater expected return than low dividend yielding stocks, sufficiently
greater to offset any tax disadvantages (Blume, 1980). Graham and Dodd ran a regression
of stock price versus dividends and retentions found that the dividend coefficient was
approximately three times the retention coefficient which implies that the investors are
willing to pay a two dollar premium per share for every dollar that is shifted from
retentions to dividends (Arditti, 1967). More recent statistical studies by Gordon (1959),
Durand (1959) and others indicate that the dividend multiplier is still several times the
retained earnings multiplier (Friend and Puckett, 1964).
4
http://amourtan.com/wp-content/uploads/2013/03/TheHighDivAdvantageStudyFUNDweb.pdf
83
There is huge amount of literature which suggests that the portfolios comprising
of high dividend yielding stocks generate higher returns in relation to low dividend
yielding stocks and overall market, such as, Arditti (1967), Litzenberger and Ramaswamy
(1979), Rozeff (1984), Levis (1989), Keppler (1991), Hodrick (1992), Brouwer et al.
(1996), Bauman et al. (1998), Fama and French (1998), Morgan and Thomas (1998),
Naranjo et al. (1998), Samarakoon (1999), Mcmanus et al. (2002), Al-Twaijry (2006),
Kyriazis and Diacogiannis (2007), Lemmon and Nguyen (2008), Safari (2009), Al-
Mwalla et al. (2010) and Khrawish (2011).
Despite the huge range of statistical results tending to confirm the existence of a
strong dividend effect, many market analysts have become increasingly unconvinced of
their validity or marginally small amount of return predictability. The debate over the
importance of dividend policy first appeared in a study by Miller and Modigliani (1961),
who concluded that in a world of perfect capital markets, the payment of dividends does
not affect the value of the firm and is therefore irrelevant. Along with it, other studies,
such as, Friend and Puckett (1964), Black and Scholes (1974), Blume (1980), Keim
(1985), Campbell and Shiller (1988), Goetzmann and Jorion (1993), Wolf (1997), Belke
and Polleit (2004), Annaert et al. (2010), Malliaropulos and Priestley (2011) also found
no evidence of predictability of stock returns through dividends.
According to Reddy and Rath (2005) “Dividend trends for a large sample of
stocks traded on Indian markets indicate that the percentage of companies paying
dividends declined from over 57 percent in 1991 to 32 percent in 2001, and that only a
few firms paid regular dividends. The profitability of firms in general appears to have
come down, and this may have caused a decline in the ability of Indian firms to pay
dividends”. It thus becomes imperative to test whether the dividend principle holds in
Indian stock market context or not.
Hence to examine the dividend rule of Graham in Indian stock market, we attempt
to test the following hypothesis:
H03 : The mean/ median market adjusted return of the stocks selected on
the basis of dividend yield at least 2/3 the yield on 91 days Treasury
bill is equal to zero.
84
In order to examine the above hypothesis, the stocks having the dividend yield
greater than two-third of yield on 91 days Treasury bill, are screened every year. Then,
the excess return over the market, generated by the stocks under study is calculated and
the significance of the returns has been examined using t-test. Table 4.3 reports the
results.
having Minimum Dividend Yield of Two-Third the AAA Bond Yield
No. of
stocks
-65.4422 -40.8539
1996 3 41.19564 - .095* 47.78007 - .285
(23.78432) (27.58584)
-17.4351 - -8.2563 -
1997 14 28.22289 .038** 29.25612 .310
(7.54288) 2.311 (7.81903) 1.056
4.4779 -6.5139
1998 24 46.51722 .472 .642 32.38989 -.985 .335
(9.49529) (6.61156)
38.3481 -.2059
1999 17 99.64561 1.587 .132 42.21586 -.020 .984
(24.16761) (10.23885)
6.2135 36.6652
2000 25 48.78198 .637 .530 25.06154 7.315 .000***
(9.75640) (5.01231)
90.9208 40.7799
2001 23 59.63713 7.312 .000*** 22.70867 8.612 .000***
(12.43520) (4.73508)
21.9479 8.5951
2002 21 35.56515 2.828 .010** 23.81438 1.654 .114
(7.76095) (5.19672)
9.4645 9.6860
2003 36 36.36478 1.562 .127 26.60439 2.184 .036**
(6.06080) (4.43406)
21.6714 -.3595
2004 44 43.88055 3.276 .002*** 23.27339 -.102 .919
(6.61524) (3.50860)
-30.7148 - -18.2446 -
2005 17 31.27102 .001*** 16.46757 .000***
(7.58434) 4.050 (3.99397) 4.568
-18.3769 - -7.1565
2006 13 24.37229 .019** 26.74518 -.965 .354
(6.75966) 2.719 (7.41778)
-3.8386 4.7814
2007 8 47.03421 - .575 18.59354 - .575
(16.62911) (6.57381)
7.5576 14.2692
2008 6 18.43684 1.004 .345 18.48722 1.891 .173
(7.52681) (7.54737)
38.6497 12.7460
2009 61 30.61138 9.861 .000*** 17.07257 5.831 .000***
(3.91939) (2.18592)
-10.2000 -5.4520
2010 13 38.87934 -.946 .363 25.11465 -.783 .449
(10.78319) (6.96555)
Across
the 17.8439 7.3762
325 52.82194 6.090 .000*** 29.41753 4.520 .000***
period (2.93003) (1.63179)
Note:
85
From Table 4.3, we notice that the number of stocks selected meeting the
principle of dividend yield varies from 3 to 61. The mean market adjusted returns of the
stocks in case of 12 months holding period have been positive or greater than zero in
1998, 1999, 2000, 2001, 2002, 2003, 2004, 2008 and 2009. However, the positive market
adjusted returns were significant only in 2001, 2002, 2004 and 2009. Also, the returns
have been significantly lesser than zero in 1996, 1997, 2005, and 2006. However, across
the period of 15 years, the stocks selected on the basis of this principle provide
significantly positive mean market adjusted return of 17.84% to investors. Similarly, the
results have been reported taking 24 months as the holding period.
When the holding period of the stocks selected on the basis of principle of
dividend yield of at least two-third the yield on 91 days Treasury bill, is extended from
12 to 24 months, the stocks show significantly negative return in only one year i.e. 2005.
In rest of the years, the negative market adjusted returns have been insignificant. Also, in
years; 2000, 2001, 2003, and 2009, the portfolio is providing significantly positive
market adjusted returns. Across the period, the principle provides the significant mean
market adjusted annualized rate of return of 7.37% (significant at 1% level of
significance) in case of 24 months holding period.
Thus, we find that the principle of dividend yield has generated significantly
higher returns than the market in Indian stock market. In spite of huge criticism of higher
dividend yields, the principle symbols its place in Indian stock market
4.2.1 P4- Discount to Tangible Book Value: The market price should be lesser
than two- third of the tangible book value per share
The price to book value ratio is the most important determinant in the value
investing world to determine whether the stock is overvalued or undervalued. Book
value per share is an accounting concept that measures what shareholders would receive
if all the firm‟s liabilities were paid off and all its assets could be sold at their balance
sheet value (Strong, 2004). It is calculated by dividing current market price per share by
the most recent 12- month period book value per share. It is important to note that Ben
Graham has focused on tangible book value in order to calculate this ratio. The intangible
assets like goodwill, trade names, patents, franchises, leaseholds etc., however have been
excluded from the calculation of book value.
86
The tangible book value per share of the common stock is found by adding up all
the tangible assets, subtracting all liabilities and stock issues ahead of the common stock
and then dividing the resultant value by the number of shares. Thus;
Tangible book value per share= (Common stock+ surplus items-intangibles)/
Number of shares outstanding or (Tangible assets-long term debt-preferred stock- current
liabilities)/ Number of shares outstanding at the year-end (Graham and Dodd, 1934).
According to Graham and Dodd (1934), “The reason behind excluding the
intangibles from book value calculation is the fact that it is not possible to estimate the
intangible constituent of a company accurately. If the present scenario is compared with
the instance a generation or more ago, one would observe a vital difference in the
valuation of intangible assets. Prior to 1920s, the intangibles were appraised on more
conservative basis than the tangibles whereas a complete reverse has taken place now,
where they are assessed more liberally than the tangibles due to greater emphasis on the
growth expectations. Thus, assumed superior expectations of the increased profits in the
future by buying undervalued securities become the major point behind appraising the
intangibles liberally.”
According to Graham (1949), “A risk adverse investor should look for the issues
which are selling in market at a price lesser than their tangible asset values. Thus, the
investor in sound asset value company, which is selling at bargain price, is owner of
interest in sound and growing business, rather than going for the stocks with glamorous
and perilous field of probable growth more influenced by the vagaries and the
fluctuations of the stock market. Also, a true investment should represent margin of
safety. Graham recommended that the stocks should be purchased when they are selling
at two-third or less of their book value. The price should be little enough to create the
considerable margin of safety. The too little margin can render the stock speculative.
Thus, sufficiently low price can turn a security of average quality into a sound investment
opportunity provided the buyer practices the sufficient diversification.”
Likewise various economic presumptions exist in favor of purchases made far
below and above the asset value. The firm which sells in market at a price higher than its
intrinsic value, earns a larger return upon its capital, which eventually attracts superior
competition and the firm that sells at lower price than the asset value, earns lesser return
87
on capital, thereby leading to removal of existing competition from the market and
restoring a normal rate of return on the investment (Graham and Dodd, 1934). Thus, in
both cases, the stock is said to be reverting to its mean. Therefore, it is advisable to select
the stocks that have greater intrinsic value than its price because the lower price and the
excellent value creates a margin of safety for the investors and when price reverts
towards its industry or economy wide mean, the investors enjoy good returns in long run.
The price to book value ratio is the cornerstone of majority of value investing
studies, such as, Stattman (1980), Rosenberg et al. (1985), Chan et al. (1991), Fama and
French (1992, 1998), Capaul et. al. (1993), Brouwer et al. (1996), Vos and Pepper (1997),
Mukherji et al. (1997), Vaidyanathan and Chava (1997), Bauman et al. (1998),
Arshanapalli et al. (1998), Dhatt et al. (1999), Doukas et al. (2001), Dimson et al. (2003),
Bird and Gerlach (2003), Ding et al. (2005) and Azzopardi (2006). However, many
studies still have not found the premium generated by price to book value ratio, such as,
Bauman and Miller (1997), Lougram (1997), Kyriazis and Diacogiannis (2007). This
could be due to the reason that the book value of equity is the accountant‟s measure of
what equity in a firm is worth and the credibility of accountants has declined over the last
few years. Nonetheless, there are many who continue to believe that accountants provide
not only a more conservative but also a more realistic measure of what equity is truly
worth than financial markets which they view as subject to irrational buying and selling.
A logical consequence of this view is that stocks that trade at a price substantially lesser
than book value are undervalued and those that trade at price greater than book value are
considered overvalued (Damodaran, 2004). We, therefore examine whether the stocks
selected on the basis of this principle helps the investors to outperform the market or not.
We thus test the following hypothesis:
H04 : The mean market adjusted return of the stocks having market price
lesser than the two- third of the tangible book value per share is equal
to zero.
In order to test this hypothesis, for each stock, the tangible book value per share
has been calculated and the stocks having market price lesser than two-third of the
tangible book value per share of the stock, are screened every year from 1996 to 2010.
Then, the significance of the market adjusted returns of the stocks meeting the above rule
has been examined in Table 4.4 for the holding period of 12 months and 24 months.
88

having Current Price Lesser than Two-Third of Its Tangible Book
Value Per Share

No. of
stocks
-62.1997 -43.7859
1996 16 39.84058 -6.245 .000*** 37.58881 -4.659 .000***
(9.96015) (9.39720)
-13.8851 -8.7654
1997 26 32.53938 -2.176 .039** 27.53941 -1.623 .117
(6.38150) (5.40092)
1.6680 -3.9585
1998 34 41.00549 .237 .814 28.73498 -.803 .428
(7.03238) (4.92801)
12.6001 3.0906
1999 42 77.28875 1.057 .297 35.07109 .571 .571
(11.92591) (5.41159)
2.7245 35.2571
2000 47 49.26872 .379 .706 25.51430 9.474 .000***
(7.18658) (3.72164)
74.7836 37.8364
2001 62 55.79036 10.555 .000*** 22.72651 13.109 .000***
(7.08538) (2.88627)
21.3101 7.3148
2002 42 32.80969 4.209 .000*** 28.73006 1.650 .107
(5.06264) (4.43314)
14.6060 21.1456
2003 46 49.79978 1.989 .053* 21.87468 6.556 .000***
(7.34258) (3.22525)
61.5693 15.3525
2004 42 62.12374 6.423 .000*** 25.30187 3.932 .000***
(9.58590) (3.90416)
-13.8871 -7.2461
2005 32 55.41123 -1.418 .166 30.02650 -1.365 .182
(9.79541) (5.30798)
4.2485 -.1543
2006 31 45.76620 .517 .609 27.83221 -.031 .976
(8.21985) (4.99881)
11.9364 8.3561
2007 26 69.94138 .870 .392 30.95910 1.376 .181
(13.71663) (6.07158)
20.4726 24.5702
2008 49 36.35448 3.942 .000*** 22.14750 7.766 .000***
(5.19350) (3.16393)
39.1761 8.4062
2009 64 36.80068 8.516 .000*** 25.53681 2.633 .011**
(4.60009) (3.19210)
-14.2897 -12.8196
2010 65 40.64863 -2.834 .006*** 25.92093 -3.987 .000***
(5.04184) (3.21510)
Across
17.1898 9.4290
the 624 57.61375 7.453 .000*** 32.30062 7.292 .000***
(2.30640) (1.29306)
period
Note:
89
From Table 4.4 we notice that the number of stocks selected on the basis of
criterion of market price lesser than the two- third of the tangible value per share and
examined under the holding period of 12 months, varies from 16 in 1996 to 65 in 2010. It
is important to note that the mean market adjusted returns yielded by the portfolio have
been greater than zero (positive) in all the years except 1996, 1997, 2005 and 2010.
However in 2001, 2002, 2003, 2004, 2008 and 2009, the returns have been significantly
greater than zero. Thus, out of 11 years period with positive market adjusted returns, the
returns have been significantly positive in 6 years. Also, across the period of 15 years, the
criterion provides market adjusted rate of return of 17.18% (significant at 1% level of
significance) to the investors.
Table 4.4 also exemplifies the results of the stocks selected on the basis of market
price lesser than two-third of the tangible book value per share and held for the period of
24 months. It is important to note that the returns so arrived have been significantly lesser
than the market only in 1996 and 2010. However, the returns have been significantly
greater than the market in 2000, 2001, 2003, 2004, 2008 and 2009. Thus, in majority of
the years, the return on the stocks selected and held for 24 months on the basis of this
principle is significantly greater than that of market returns. Thus, across the period of 15
years, the criterion yields the annualized market adjusted rate of return of 9.42%
(significant at 1% level of significance) to the investors.
4.2.1 P5- Net Current Asset Value (NCAV): The market price should be lesser
than two-third of the net current asset value per share
According to Graham, net current asset value (NCAV) is the basic key of the
liquidating worth of the enterprise. Liquidating value of the enterprise is the money that
the owners could get out of it when the company is liquidated. It is calculated as under:
Net current asset value= Current assets- (long term as well as short term
liabilities+ preferred stock). The long term assets are not included in its calculation
(Graham, 1934).
This strategy is considered to be the subset of the price to book value strategy.
However, the book value can contain illiquid assets or insubstantial assets that do not
90
hold value or cannot be liquidated without continuing operations. It could be possible that
the long term assets have far higher book values than true economic value. The NCAV
strategy takes into account current assets only and excludes the assets that are difficult to
value (Guenster, 2009).
According to this principle, a stock should be bought when its market price is
lesser than two- third of its net current asset value. This would mean that the buyer would
pay nothing at all for the fixed assets, such as, land, buildings, machinery etc., or any
goodwill items that might exist (Graham, 1949). Graham found that companies satisfying
this principle were often priced at significant discounts to estimates of the value that
shareholders could receive in an actual sale or liquidation of the entire enterprise. Thus,
the principle not only protects capital from significant lasting loss but also yields a
portfolio of stocks with excellent prospects for advance in price (Xiao and Arnold, 2008).
Graham and Dodd (1934), stated that “there could be two way implication of this
principle. The first is that the stocks selling below the net current asset value are
considered too inexpensive and offer a lucrative margin of safety to investors. On the
other hand, such a low price could impel the stockholders to raise the question whether or
not it is in the interest to continue the business. This is due to the fact that the issue
selling below the liquidating value is the signal that mistaken policies are being followed
which would impel the management to take all the proper steps to correct the disparity
between market quotation and the intrinsic value.” According to Guenster (2009), “The
logical phenomenon behind the profitability of this principle is as under:
Due to the mean reversion tendency of the stocks, the earnings power of the
securities that have low earnings currently would eventually increase due to the
fact that a business or industry with a record of low profitability can improve its
business when the existing competition is reduced from the field. Another reason
for an increase in earnings could be due to a change in operating policy of the
business i.e. loss-making businesses can be closed down or the management can
change the strategy of the business. This can either be done voluntarily or they
can be forced to change their policy by the shareholders. When earnings go up
the return on assets will increase thereby often increasing the share price.
91
The second possibility is that such firms could be sold to the rivals who are
capable of utilizing the assets more efficiently. In this situation, the rivals would
pay at least the liquidation value of the assets.
The third option available in case of such firms is the discontinuance of the
business. Since the firm is trading at a price lesser than its net current assets value,
the liquidation would release at least the value of tangible current assets. Thus,
liquidation of such firm will be profitable.”
Thus, when one purchases the group of stocks meeting this criterion, there is a
strong potential that the forces counteracting the trend would unlock the discount present
in stocks and will increase the value of stocks leading to satisfactory returns to the
investors (Graham and Dodd, 1934; Guenster, 2009).
Graham developed and tested this criterion between 1930 and 1932, and it was
used extensively in the operations of the Graham-Newman Corporation through 1956.
Graham reported that issues selected on the basis of the rule earned, on average, about 20
per cent per year over a 30-year period. After the mid-1950s, however, bargain issues
became relatively scarce. Some issues became available again during the early 1970s,
following the market declines of the late 1960s, and became abundant after the 1973-74
bear market (Oppenheimer, 1986). Also many practitioners believe that it is very difficult
to find such bargain issues in present economic environment. Thus, we examine if there
are such bargains existing in Indian stock market and examine their performance through
following hypothesis:
H05 : The mean/ median market adjusted return of the stocks having
market price lesser than the two- third of the net current asset value
per share is equal to zero.
In order to test this hypothesis, for each stock, the net current asset value per share
has been calculated and the stocks having market price lesser than two-third of the net
current asset value per share of the stock, are screened on 30th June every year from 1996
to 2010. Then, the statistical significance of the market adjusted returns of the stocks
meeting the above rule has been examined and the results are reported in Table 4.5 in
case of holding period of 12 months, 24 months.
92

having Current Price Lesser than Two-Third of Its Net Current Asset
Value Per Share
No. of
stocks
1996 0 - - - - - - - -
1997 1 61.5927 - - - 13.5207 - - -
1998 1 -30.5179 - - - 7.6757 - - -
60.3800 25.7708
1999 5 45.63260 - .043** 19.07857 - .043**
(20.40752) (8.53219)
2.6172 40.2771
2000 10 54.25657 .153 .882 31.17182 4.086 .003***
(17.15743) (9.85740)
118.4744 51.9700
2001 9 85.11620 4.176 .003*** 32.08057 4.860 .001***
(28.37207) (10.69352)
8.2411 5.4622
2002 5 24.40634 - .592 36.94332 - .500
(10.91485) (16.52156)
27.2762 34.8248
2003 5 31.41704 - .138 19.18249 - .043**
(14.05013) (8.57867)
77.3113 21.7296
2004 8 68.15783 3.208 .015** 26.28536 2.338 .052*
(24.09743) (9.29328)
13.2796 5.2909
2005 6 68.94510 - .753 33.83049 - .600
(28.14672) (13.81124)
10.8652 18.6709
2006 6 54.12487 - .753 28.33705 - .173
(22.09639) (11.56855)
65.8238 25.3630
2007 5 97.29697 - .241 47.29593 - .139
(43.51253) (21.15138)
21.7703 18.0418
2008 10 44.03436 1.563 .205 30.84026 1.850 .097*
(13.92489) (9.75255)
53.2555 11.5725
2009 11 49.05963 3.600 .005*** 35.09205 1.094 .300
(14.79204) (10.58065)
-12.8825 -9.8378
2010 9 39.66573 -.974 .358 19.22546 -1.53 .163
(13.22191) (6.40849)
Across
37.1764 20.7207
the 91 66.52767 5.331 .000*** 33.13380 5.966 .000***
(6.97399) (3.47337)
period
Note:
93
Table 4.5 shows that the stocks meeting the principle of having market price
lesser than the two-third of the net current asset value could not be found in year 1996.
Further, in year 1997 and 1998, only one such stock could be found each year. Further, in
rest of years the stocks arrived at meeting this principle ranged from 5 to 11. It is further
interesting to note that the mean market adjusted returns in case of 12 months holding
period have been positive in all the years except 1988 and 2010. Also, in 1999, 2001,
2004 and 2009, the mean market adjusted returns have been significantly positive. Thus,
across the period of 15 years, the stocks selected on the basis of this principle provide
mean market adjusted return of 37.18% which is significant at 1% level of significance.
Table 4.5 also reports the results regarding the performance of the stocks selected
on the basis of market price lesser than two-third of the net current asset value per share
and held for the period of 24 months. It is important to note that the annualized rate of
return so arrived at is greater than the market in all the years except 2010. Also, in 2010
the portfolio generates statistically insignificant annualized return. Rest, all the years
show the positive mean market adjusted rate of return. Moreover, in 2000 and 2001, the
returns have been significant at 1% level of significance and in 1999 and 2003; the
returns have been significant at 5% level of significance. In 2004 and 2008, the returns so
attained have been significant at 10% level of significance. Further, across the period of
the study, the annualized market adjusted rate of return of 20.7% generated by NCAV
stocks has been significant at 1% level of significance. Thus, this strategy of buying
stocks has yielded significantly high rate of return to investors in both the holding periods
in maximum number of years and across the period. These results are consistent with the
different studies that have examined the net current asset value rule and obtained
significant mean market adjusted returns over the study period (Oppenheimer, 1986;
Lauterbach and Vu, 1993; Xiao and Arnold, 2008; Guenster, 2009; Mohanty et al., 2012).
Hence, the strategy has been profitable to the investors in Indian stock market.
4.2.1 P6: Debt to Equity: Total debt lesser than the book value
The debt means the total amount of money borrowed by a company from outside
parties. It includes both, short term as well as long term liabilities owned by a company to
94
outside parties. Book value of firm means the excess of assets over total debt and
preferred stock. This principle states that the debt should not exceed the book value or the
debt to equity ratio should be lesser than 1. This principle measures the financial strength
of the company. The lesser debt in relation to book value does not always mean that bank
debt is a bad sign in itself. The use of a reasonable amount of bank credit particularly for
seasonal needs is not only legitimate but even desirable (Graham and Dodd, 1934).
However, the higher amount of debt can put financial problems on the company at the
time of depression. If the firm has higher debt in relation to book value, then earnings
would be used to pay off debt and the excessive debt could also lead to bankruptcy. Thus,
the companies should have sound asset base to pay off its debt.
However, the traditional notion of low debt phenomenon by Graham has changed
in the present economic environment. Leverage, for decades, has the important place in
financial structure of the company. Financial leverage is the share of capital financed
through debt as opposed to equity. Higher leverage implies higher proportion of debt in
the capital structure of the company. The impact of leverage on the value of the firm is
the controversial issue. Some authors have proposed its positive impact in enhancing
overall returns and other have opined the negative impact of leverage on firm‟s value.
Modigliani and Miller (1958) are first to form the basis for modern opinion on capital
structure. They have asserted that in the absence of taxes and transaction cost, the capital
structure is not having any influence on the firm‟s overall value. Thus, it is irrelevant if
the capital is financed by debt or equity. They stated that the rational firm is one which
will tend to maximize its profits and market value through the mix of capital structure.
According to Modigliani and Miller (1958), “A rational firm is one which will tend to
push investment to the point where the marginal yield on physical assets is equal to the
market rate of interest on bonds or debt. Hence, a physical asset is worth acquiring if it
will increase the net profit of the owners of the firm. But net profit will increase only if
the expected rate of return on the asset exceeds the rate of interest. The risk due to
increased debt would be offset by the increased profitability and the market value of the
firm”. Thus the value of the firm remains same, regardless of whether the funds are
95
acquired through debt instruments or through new issues of common stock. Further,
taking the impact of taxes in their study, Modigliani and Miller (1963) found that the debt
increases the riskiness of the stock and as a result, equity shareholders demand a higher
return. The interest expense deduction increases the firm value and reduces the leverage
related risk premium demanded by equity investors. Hence, the value of levered and
unlevered firm remains same.
Some authors stated that debt equity ratio is just a proxy for risk and financial
risk premium is positively related to firm‟s leverage and have found the positive relation
between leverage and stock returns. The studies like Hamada (1972), Bhandari (1988),
Mukherji et al. (1997), Fama and French (2002), Lasher (2003), Ding et al. (2005),
Dhaliwal et al. (2006), Ward and Price (2006), Sharma (2006), Tripathi (2009) etc. show
that returns increase with the increase in leverage. Ross (1977) and Heinkel (1982) also
asserted that if the manager is going for leverage, then it is an indication that the
company has investment opportunities and signal that it is able to service its debt. Marsh
(1982) and Ooi (1999) also asserted that high growth firms and the profitable firms will
go for more debt due to tax benefit provided by debt and they have a low risk of
bankruptcy. It therefore becomes imperative to examine if the original idea of low debt in
capital structure can help to augment returns or not. Therefore, we test the following
hypothesis
H06 : The mean market adjusted return of the stocks having total debt
lesser than the book value is equal to zero.
In order to test the above hypothesis, total debt (sum of short term and long term
debt) as well as book value of all the stocks in the sample is calculated and the stocks
having lesser total debt than the book value are selected. Further, of the selected stocks,
the returns have been observed of the period of 12 months as well as 24 months. Table
4.6 reports the results of one sample t-test used to determine the significance of returns of
the selected stocks in case of 12 months, 24 months holding period.
96

Having Total Debt Lesser than the Book Value
No. of
stocks
-29.5673 -29.5659
1996 62 48.66735 -4.784 .000*** 39.94269 -5.828 .000***
(6.18076) (5.07273)
-10.3719 -4.0968
1997 59 53.12443 -1.500 .139 32.08759 -.981 .331
(6.91621) (4.17745)
14.2121 -1.7519
1998 58 44.24114 2.447 .018** 31.99072 -.417 .678
(5.80915) (4.20059)
-.3513 -6.4241
1999 69 76.22965 -.038 .970 39.43112 -1.353 .180
(9.17697) (4.74695)
-2.6670 23.9105
2000 80 46.89772 -.509 .612 27.89093 7.668 .000***
(5.24332) (3.11830)
55.1034 28.5212
2001 83 56.20739 8.931 .000*** 25.19161 10.315 .000***
(6.16956) (2.76514)
17.8845 3.7986
2002 82 29.59013 5.473 .000*** 24.67128 1.394 .167
(3.26768) (2.72449)
1.5068 7.8332
2003 89 44.35433 .320 .749 23.78117 3.107 .003***
(4.70155) (2.52080)
25.4063 3.8542
2004 111 50.70735 5.279 .000*** 23.91441 1.698 .092*
(4.81293) (2.26986)
-8.5567 -7.2384
2005 130 44.29923 -2.202 .029** 23.14505 -3.566 .001***
(3.88530) (2.02995)
-2.5500 -7.3403
2006 144 36.36181 -.842 .401 25.22921 -3.491 .001***
(3.03015) (2.10243)
-9.6991 4.5301
2007 165 53.71436 -2.319 .022** 27.02378 2.153 .033**
(4.18166) (2.10380)
25.8503 23.4680
2008 192 41.08774 8.718 .000*** 19.92937 16.317 .000***
(2.96525) (1.43828)
33.4862 5.8825
2009 190 36.49199 12.649 .000*** 24.81762 3.267 .001***
(2.64741) (1.80046)
-12.5394 -8.4755
2010 217 34.87593 -5.296 .000*** 26.43810 -4.722 .000***
(2.36753) (1.79474)
Across
7.3140 3.2529
the 1731 49.61483 6.133 .000*** 29.70975 4.555 .000***
(1.19251) (.71409)
period
Note:
97
Table 4.6 shows that the number of stocks meeting the principle of having lesser
debt in relation to book value (P6) ranges from 58 to 217 across the period of study. Out
of the total period of the study, the mean market adjusted returns have been lesser than
the market in 8 years, that is, 1996, 1997, 1999, 2000, 2005, 2006, 2007 and 2010. In
rest of the years, i.e. 1998, 2001, 2002, 2004, 2008 and 2009, the rate of return has been
significantly larger than the market in case of 12 months holding period. Only in 2003,
the positive market adjusted returns attained have been insignificant due to larger
volatility of the stocks. Overall, in the entire period of 15 years, the criterion generates
the market adjusted rate of return of 7.3% (significant at 1% level of significance) to the
investors.
Also apparent from Table 4.6, when the holding period of these stocks is extended
from 12 to 24 months, the number of years in which the criterion generates negative
returns decreases from 8 to 7 years. Also, (except 2010) in rest of the years, the mean
market adjusted return of the stocks has been significantly larger than that of market.
Across the period of 15 years, the criterion generates mean market adjusted annualized
rate of return of 3.25%, which is significant at 1% level of significance.
4.2.1 P7- Current Ratio: Current ratio of a company should be greater than two
Current ratio determines how many dollars in assets are likely to be converted to
cash in one year in order to pay off the debts that become payable during the same year.
Working capital (excess of current assets over current liabilities) decision affects both
liquidity and profitability. Excess of investment in working capital may result in low
profitability and lower investment may result in poor liquidity. Management need to
trade-off between liquidity and profitability to maximize shareholders wealth (Chary et
al., 2011). Graham (1934) believed that current assets i.e. cash, marketable securities,
receivables, and merchandise inventory etc. are either directly equivalent to cash, or are
likely t o be converted into cash, through sale or collection, in the ordinary course of
business. An industrial enterprise should therefore possess a considerable amount of
current assets over its current liabilities in order to conduct its operations successfully.
98
Working capital management involves planning and controlling current assets and current
liabilities in a manner that eliminates the risk of inability to meet short term obligations
on the one hand and avoid excessive investment in these assets on the other hand (Eljelly,
2004). Generally, it is assumed that the current liabilities must be met by current assets.
The maturity date of current assets should coincide with maturity date of current
liabilities (maximum maturity date is one year). Lack of coincidence between maturity
date of current assets and the maturity date of current liabilities leads to liquidity
problems for the firms (Alavinasab and Davoudi, 2013). Every company needs to
maintain adequate working capital so as to meet requirements of different stakeholders,
for example, timely payment of its obligations to suppliers, timely payment of salaries
and wages to its employees, etc.
It is not viable to fix explicit minimum requirements for any one of these factors,
especially, because the normal working capital situation varies generally with different
types of venture. It is generally held that current assets should be at least double the
current liabilities, in order to have a sizeable cushion of working capital that on average
should uphold a company through its hard times and a smaller ratio would
unquestionably call for further examination (Graham, 1949). Thus the following
hypothesis is tested:
H07 : The mean market adjusted return of stocks having current ratio
greater than 2 is equal to zero.
In order to examine this hypothesis, the current ratio of all the stocks is computed
and the stocks having current ratio greater than 2 are selected to examine their market
performance. Table 4.7 reports the results.
99

Having Current Ratio Greater than Two
No. of
stocks
-18.5506 - -20.8788 -
1996 27 42.32259 .031** 40.27899 .012**
(8.14499) 2.278 (7.75170) 2.693
-15.4705 - -8.0538 -
1997 32 63.36048 .177 37.04924 .228
(11.20066) 1.381 (6.54944) 1.230
17.0915 1.7667
1998 32 47.48818 2.036 .050* 34.72386 .288 .775
(8.39480) (6.13837)
18.1893 .4938
1999 30 84.11668 1.184 .246 38.78221 .070 .945
(15.35753) (7.08063)
-9.7883 - 21.1696
2000 41 41.51814 .139 29.56068 4.586 .000***
(6.48404) 1.510 (4.61660)
56.2797 30.0443
2001 53 50.91934 8.046 .000*** 24.03015 9.102 .000***
(6.99431) (3.30080)
20.9452 6.1837
2002 41 31.33026 4.281 .000*** 24.32020 1.628 .111
(4.89296) (3.79818)
7.0916 6.0335
2003 41 44.50667 1.020 .314 20.64407 1.871 .069*
(6.95077) (3.22406)
15.1846 -3.0570 -
2004 43 31.45790 3.165 .003*** 19.50420 .310
(4.79728) (2.97436) 1.028
-22.7310 - -13.7425 -
2005 53 41.55063 .000*** 24.24578 .000***
(5.70742) 3.983 (3.33041) 4.126
-10.5655 - -5.9124 -
2006 64 35.64769 .021** 26.28231 .077*
(4.45596) 2.371 (3.28529) 1.800
-1.5703 2.4187
2007 87 63.95759 -.229 .819 29.45116 .766 .446
(6.85697) (3.15750)
16.3869 19.3628
2008 103 47.20277 3.523 .001*** 22.70046 8.657 .000***
(4.65103) (2.23674)
33.5726 7.4384
2009 94 34.91293 9.323 .000*** 22.58671 3.193 .002***
(3.60099) (2.32964)
-14.3488 - -12.7268 -
2010 112 41.40111 .000*** 32.78466 .000***
(3.91204) 3.668 (3.09786) 4.108
Across
6.6819 2.8740
the 853 51.43231 3.794 .000*** 30.77035 2.728 .007***
(1.76101) (1.05356)
period
Note:
From Table 4.7 we note that the number of stocks meeting the said principle
range from 27 in 1996 to 112 in 2010. Looking at the market adjusted return of these
stocks, we note that the criterion generates negative average market adjusted returns in 7
100
years i.e. 1996, 1997, 2000, 2005, 2006, 2007 and 2010. However the negative returns
have been significant only in year 1996, 2006 and 2010. The stocks generate positive
market adjusted returns in 1998, 1999, 2001, 2002, 2003, 2004, 2008 and 2009. Out of
these years, the positive market adjusted returns have been significant in 2001, 2002,
2004, 2008 and 2009. Across the period of 15 years, the stocks having current ratio
greater than 2 generates significantly positive market adjusted rate of return of 6.68% to
investors.
From Table 4.7, we further notice that in case of 24 months holding period, in 9
out of 15 year period, the said portfolio provides mean annualized returns greater than
that of market returns. Out of these 9 years, in 5 years, i.e., 2000, 2001, 2003, 2008 and
2009, the average annualized returns have been significantly greater than that of market
returns. However, portfolio yielded negative market adjusted returns in 1996, 1997, 2004,
2005, 2006 and 2010. Also, in 1996, 2005, 2006 and 2010, the returns are significantly
lesser than that of market. However, across the period of 15 years, the mean market
adjusted annualized rate of return of 2.87% is statistically significant at 1% level of
significance.
4.2.1 P8: Debt to NCAV: The total debt should be lesser than twice the net current
asset value
This principle has also been formed to protect the business in case of higher
amount of debt. According to Graham (1949), “Whenever a bond is completely covered
by net current assets, its safety is thereby guaranteed. Thus, the net current assets of the
company should be sufficient enough to provide complete coat to the total debt of the
company. However, certain types of industrials e.g., baking, ice and restaurant concerns
normally require a relatively small amount of working capital in relation to total assets
and business. For such businesses, 100% net current asset coverage requirement for
bonds would be over stringent”. Thus, the total debt of a company should be lesser than
twice the net current asset value of the company. Therefore, we examine the following
hypothesis
H08 : The market adjusted returns of stocks having total debt lesser than
twice the net current asset is equal to zero
101
In order to test this hypothesis, all the stocks having total debt lesser than twice its
net current asset are selected, every year from 1996 to 2010 and their market adjusted
returns are tested using one sample t-test. Table 4.8 reports the results.
Having Total Debt Lesser than Twice the Net Current Asset Value
No. of
Year Mean T- P- Mean Std. T- P-
stocks Std. Dev.
(Annual) Value Value (Annualized) Dev. Value Value
-15.6099 -24.1398
1996 28 49.07792 -1.683 .104 46.30851 -2.758 .010**
(9.27485) (8.75149)
-.0547 3.2324
1997 20 75.57623 -.003 .997 36.68619 .394 .698
(16.89936) (8.20328)
27.9908 4.2411
1998 22 47.40787 2.769 .011** 33.26235 .598 .556
(10.10739) (7.09156)
-3.2695 -7.1302
1999 29 70.13283 -.251 .804 39.95821 -.961 .345
(13.02334) (7.42005)
-.9459 20.6106
2000 37 45.34187 -.127 .900 32.05562 3.911 .000***
(7.45416) (5.26991)
52.7936 25.8710
2001 40 64.34233 5.189 .000*** 30.23062 5.412 .000***
(10.17342) (4.77988)
15.3721 -.0666
2002 39 28.49365 3.369 .002*** 22.15963 -.019 .985
(4.56264) (3.54838)
-4.9071 2.1285
2003 47 36.47039 -.922 .361 23.73981 .615 .542
(5.31975) (3.46281)
19.7558 2.3491
2004 62 45.29368 3.434 .001*** 23.70648 .780 .438
(5.75230) (3.01073)
-3.1102 -5.3639
2005 72 40.59348 -.650 .518 20.70424 -2.198 .031**
(4.78399) (2.44002)
-2.9141 -4.7018
2006 82 33.75552 -.782 .437 21.56363 -1.974 .052*
(3.72767) (2.38131)
-5.9622 7.6082
2007 99 54.32005 -1.092 .277 25.40942 2.979 .004***
(5.45937) (2.55374)
28.3017 24.6604
2008 111 36.50637 8.168 .000*** 20.19384 12.866 .000***
(3.46503) (1.91671)
35.5909 6.9113
2009 109 37.47421 9.916 .000*** 25.24400 2.858 .005***
(3.58938) (2.41794)
-9.4542 -5.8766
2010 117 33.77317 -3.028 .003*** 24.69200 -2.574 .011**
(3.12233) (2.28278)
Across
9.4255 4.4861
the 914 47.63569 5.982 .000*** 28.62497 4.738 .000***
(1.57565) (.94683)
period
Note:
102
As evident from Table 4.8, the number of stocks, meeting the criterion of having
total debt lesser as twice the net current assets, range from 20 to 117 across the period of
15 years. It is also pertinent to note that in 9 years (i.e. 1996, 1997, 1999, 2000, 2003,
2005, 2006, 2007 and 2010), the portfolio is generating lesser returns than that of market
returns. Out of these 9 years, only in 2010, the market adjusted returns have been
significantly negative. Further in all of the remaining 6 years i.e. 1998, 2001, 2002, 2004,
2008 and 2009, the returns have been significantly greater than that of market returns.
Thus, across the period of 15 years, the portfolio has generated market adjusted return of
9.425% (significant at 1% level of significance).
From Table 4.8 we further notice that, when the holding period of stocks having
lesser debt than twice the net current assets have been extended from 12 months to 24
months, the number of years in which the said portfolio yielded negative returns
decreased from 9 to 6 years (1996, 1999, 2002, 2005, 2006, 2010). Out of these 6 years,
the negative return is significant only in 1996, 2005, 2006 and 2010. However, in 9 out of
15 years, the mean market adjusted return has been positive and significantly positive in
2000, 2001, 2007, 2008 and 2009. Across the period of 15 years, the mean annualized
market adjusted rate of return of the stocks is 4.48% significant at 1% level of
significance.
4.2.1 P9: Earnings Growth: Earnings should have shown a compounded annual
growth of at least 7% in last ten years
This principle goes against the concept of value stocks and favors the purchase of
stocks having sound record of past earnings (generally called as glamour stocks). Thus
the two strategies; value and growth, instead of being mutually exclusive to each other,
can complement each other in enhancing the returns of the investors. Stratifying style
portfolios on rate of change in earnings indicates the efficacy of using information on
momentum to further enhance portfolio returns (Ahmed and Nanda, 2001). This principle
is concerned with the utility of the past record as an indicator of future earnings. The
record must cover a number of years, first, because a continued or repeated performance
is always more impressive than a single occurrence and secondly, the average of a fairly
103
long period of time tend to absorb and equalize the distorting influences of the business
cycle (Graham and Dodd, 1934). The strategy aimed at buying growth stocks extrapolates
past earnings growth too far into future assuming that the trend in the growth would
continue into future as the market overreacts to good or bad news. Thus, this principle
equates good investment with well run company irrespective of its market price
(Lakonishok et al., 1994).
Graham and Dodd (1934) believed that a continuous increase in profits proved
that the company was on the upgrade and promised still better results in the future than
had been accomplished to date. Conversely, if the earnings had declined or even
remained stationary during a prosperous period, the future must be thought unpromising,
and the issue was certainly to be avoided. Graham (1949) proposed that the stock that has
shown an compounded annual growth rate of at least 7% in last 10 years should be
bought by an investor. On account of limitation of past 10 years information availability
on earnings, the 5-year growth in the earnings has been calculated for the last two
principles as advocated in Lakonishok et al. (1994). Bauman and Miller (1997) have
examined 4 year growth rate and Ahmed and Nanda (2001) have used 2 year growth rate.
Vanstone et al. (2004) have, however, eliminated this principle on account of requirement
of past 10 year information while examining Graham‟s stock selection strategy.
Therefore, in order to examine this principle, the time period of 5 years have been
considered. The following hypothesis is examined:
H09 : The mean market adjusted return of stocks that have shown a
compounded annual growth rate (CAGR) of 7% or more in last 5
years is equal to zero.
Table 4.9 reports the results of testing the above hypothesis.
104

having CAGR of 7% or more in Last Five Years.
No. of
Year Mean Std. T- P- Mean Std. T- P-
stocks
(Annual) Dev. Value Value (Annualized) Dev. Value Value
-24.5019 -26.3391
1996 37 42.21678 -3.530 .001*** 39.45499 -4.061 .000***
(6.94040) (6.48636)
-8.1493 -1.7525
1997 31 65.14235 -.697 .491 36.34864 -.268 .790
(11.69991) (6.52841)
18.1884 1.1401
1998 39 43.44095 2.615 .013** 33.51440 .212 .833
(6.95612) (5.36660)
3.4280 -10.0719
1999 37 88.27214 .236 .815 42.66526 -1.436 .160
(14.51185) (7.01413)
-4.7339 19.4719
2000 38 44.34464 -.658 .515 27.41696 4.378 .000***
(7.19365) (4.44762)
60.1957 30.2162
2001 31 60.28344 5.560 .000*** 29.39534 5.723 .000***
(10.82723) (5.27956)
24.7839 4.3251
2002 27 31.99984 4.024 .000*** 27.13079 .828 .415
(6.15837) (5.22132)
1.4931 4.6986
2003 43 37.78827 .259 .797 22.38391 1.376 .176
(5.76266) (3.41351)
27.0261 3.6881
2004 56 43.73982 4.624 .000*** 22.89531 1.205 .233
(5.84498) (3.05951)
-4.7050 -2.6072
2005 70 44.84962 -.878 .383 20.98461 -1.039 .302
(5.36055) (2.50814)
4.8790 -.9268
2006 79 38.62708 1.123 .265 24.58628 -.335 .738
(4.34589) (2.76617)
-4.9242 5.7094
2007 87 58.33436 -.787 .433 25.67265 2.074 .041**
(6.25410) (2.75240)
18.8171 20.4728
2008 98 34.26999 5.436 .000*** 18.67434 10.853 .000***
(3.46179) (1.88639)
32.3188 7.0625
2009 93 31.80233 9.800 .000*** 22.55052 3.020 .003***
(3.29775) (2.33838)
-11.2879 -6.6908
2010 118 36.20417 -3.387 .001*** 26.79389 -2.713 .008***
(3.33286) (2.46658)
Across
7.6419 3.3479
the 884 49.16748 4.621 .000*** 29.03555 3.428 .001***
(1.65368) (.97657)
period
Note:
As pertinent from Table 4.9, the market adjusted returns of the stocks selected on
the basis of this principle has been positive in 9 (1998, 1999, 2001, 2002, 2003, 2004,
2006, 2008 and 2009) out of 15 years period followed by significantly positive in 1998,
2001, 2002, 2004, 2008 and 2009. However in rest of the years, the returns have been
lesser than the market, followed by significantly negative returns in 1996 and 2010.
Across the period of 15 years, the portfolio of stocks having good record of growth in
105
earnings has provided significantly larger return than the market in case of 12 months
holding period.
When the holding period of the stocks meeting the principle of earnings growth
has been extended from 12 months to 24 months, the portfolio continues to generate
significantly negative returns in 1996 and 2010. Further in rest of the years i.e. 1997,
1999, 2005 and 2006, the portfolio generates negative but statistically insignificant
market adjusted returns. Moreover, the market adjusted returns of the stocks selected on
the basis of this principle has been positive in 9 (1998, 2000, 2001, 2002, 2003, 2004,
2007, 2008 and 2009) out of 15 years period followed by significantly positive in 2000,
2001, 2007, 2008 and 2009. Thus, across the period of study, the portfolio of stocks
having good record of growth in earnings has provided an annualized market adjusted
return of 3.34 % (significant at 1% level of significance) in case of 24 months holding
period.
4.2.1 P10: Earnings Stability: There should be no more than two declines of 5 per
cent or more in year-end earnings over the most recent ten years.
When an investor invests in equity, it is exposed to the risk that firm‟s underlying
business may go through rough times and that the market price of the stock would reveal
these downswings. However, if the earnings are stable, stock in the firm becomes safer
and potentially a better investment. Securities of the companies with stable earnings yield
lesser volatile returns for its investors. Thus, the stocks having stability in earnings are
less likely to shake markets with earnings announcements that surprise investors. The
resulting price stability should make the returns on these stocks much more predictable
than returns on rest of the market, especially if the firm takes the advantage of its more
stable earnings to pay larger dividends every year (Damodaran, 2004). The last principle
of Benjamin Graham is related to the stability in per share earnings of the company. It is
not only the growth in earnings which is important, but the growth should have been
followed by lesser swings in per share earnings of the company. Hence Graham reported
that there should be no more than 2 declines of 5% or more in per share earnings in last
10 years (taken 5 years in the present study)
106
H010 : The market adjusted returns of the stocks having stability in its per
share earnings in past 5 years is equal to zero
In order to test this hypothesis, the changes in the growth rate of the earnings for
all stocks are calculated taking previous 5 years data. The stocks showing the consistent
growth i.e. not more than 2 declines of 5% or more in earnings per share are screened for
examination of their returns. Table 4.10 provides the results of the one sample t-test
employed to examine the significance of returns in case of 12 months, 24 months holding
period.
having Stability in Earnings over Past 5 Years
No. of
Year Mean Std. T- P- Mean T- P-
stocks Std. Dev.
(Annual) Dev. Value Value (Annualized) Value Value
-27.0015 - -29.6815
1996 22 51.20297 .022** 33.07933 -4.209 .000***
(10.91651) 2.473 (7.05254)
-4.5607 -1.8325
1997 23 71.22646 -.307 .762 37.05872 -.237 .815
(14.85174) (7.72728)
11.2483 .1693
1998 33 42.48350 1.521 .138 33.77506 .029 .977
(7.39543) (5.87948)
11.3658 -3.4202
1999 32 91.43553 .703 .487 41.82406 -.463 .647
(16.16367) (7.39352)
1.2474 22.6926
2000 33 45.73626 .157 .876 22.52735 5.787 .000***
(7.96166) (3.92151)
49.8566 25.6786
2001 28 57.75587 4.568 .000*** 29.52427 4.602 .000***
(10.91483) (5.57956)
21.3222 .8338
2002 22 35.96322 2.781 .011** 30.35756 .129 .899
(7.66738) (6.47225)
-5.1491 4.7249
2003 27 37.13058 -.721 .478 20.90786 1.174 .251
(7.14578) (4.02372)
31.2743 8.7894
2004 37 46.21831 4.116 .000*** 20.58660 2.597 .014**
(7.59824) (3.38442)
-2.0174 -4.0206
2005 48 39.66923 -.352 .726 19.69031 -1.415 .164
(5.72576) (2.84205)
3.5397 -3.6561
2006 58 37.31784 .722 .473 26.07064 -1.068 .290
(4.90007) (3.42324)
-6.2552 4.9081
2007 75 63.14468 -.858 .394 27.13415 1.566 .122
(7.29132) (3.13318)
22.2904 20.9322
2008 88 35.56466 5.880 .000*** 17.54584 11.191 .000***
(3.79121) (1.87039)
27.6127 3.9804
2009 78 35.50592 6.868 .000*** 25.19682 1.395 .167
(4.02025) (2.85298)
-4.1132 - -4.7475
2010 77 33.20339 .280 28.95821 -1.439 .154
(3.78388) 1.087 (3.30009)
Across
9.3954 4.2438
the 681 50.13847 4.890 .000*** 29.00083 3.819 .000***
(1.92131) (1.11131)
period
Note:
107
As apparent from Table 4.10, the stocks meeting the criterion range from 22 to 88
across the period of 15 years. Further, the stocks so arrived in the portfolio provide
positive mean market adjusted returns in 9 years out of 15 years period. Further, the
returns have been significant in 2001, 2002, 2004, 2008 and 2009. However, only in year
1996 the criterion provides significantly negative market adjusted returns. Nonetheless,
across the period of 15 years, the portfolio provides significantly positive market adjusted
return of 9.39% in case of 12 months holding period
Table 4.10 also shows that, when the holding period of the stocks having stability
in their per share earnings, is extended to 24 months, the stocks provide positive mean
market adjusted returns in 1998, 2000, 2001, 2002, 2003, 2004, 2007, 2008 and 2009.
Thus out of 15 year period, in 9 years, the mean market adjusted return has been positive
and in year 2000, 2001, 2004 and 2008 the returns have been significantly larger than the
market returns. It is further important to note that the returns have been significantly
lesser than the market only in 1996. On the whole, the principle provides mean market
adjusted return of 4.24% (significant at 1% level of significance) across the period of 15
years in case of 24 months holding period.
The above discussion shows that the applicability of the Graham‟s principles
cannot be ignored in the present scenario in Indian stock market. Even though the mean
market adjusted returns of the stocks meeting different principles have declined with the
extension of holding period from 12 months to 24 months, their deviation around mean
(i.e. standard deviation) has also shown a decline across the study period. In a nutshell,
across the period of 15 years, all the principles are generating significantly higher returns
than the market in 12 and 24 months holding periods.
108
Table 4.11: Description of the Variables Discussed in Benjamin Graham’s

StockSelection Criteria
S. Variable Description
No.
1. Earnings yield It is the inverse of price to earnings ratio. It is calculated by dividing the
most recent 12-month period earnings per share by the current market price
per share.
2. Price to earnings Current market price per share divided by the most recent 12-month period
ratio earnings per share.
3. Dividend yield Dividend per share at the end of the year divided by the current market price
per share.
4. Tangible book (Common stock+ surplus items-intangibles) divided by number of shares
value per share outstanding or (Tangible assets-long term debt-preferred stock- current
liabilities) divided by number of shares outstanding at the year end.
5. Net Current asset Current assets- (long term as well as short term liabilities+ preferred stock)
value per share divided by number of common shares outstanding.
6. Current ratio Current assets divided by the current liabilities
7. Total debt Sum of long term as well as short term liabilities
8. Book value Total assets- total debt- preferred stock
9. Earnings growth It has been calculated the same way compounded annual growth rate is
rate computed:
CAGR t0 , tn V tn / V t0 ^ 1/ tn t0 1
Where, V t0 = start value; V tn = finish value
4.2.2 Testing the Difference between Fulfillment or Non Fulfillment of the

Principles amongst the Total Sample of Stocks
The further discussion will help us to know that out of total sample of stocks
under study, the fulfillment or non fulfillment of particular principle by all the sample
stocks causes any difference in the stocks returns or not. In order to judge the significance
of difference, independent sample t-test has been employed. Out of the entire sample of
the study i.e. 1909 stocks across the period of 15 years, the stocks have been classified
into two groups- one group consisting of stocks that fulfill the particular principle and
another group consisting of stocks which do not fulfill the said principle. The following
hypothesis is tested:
H011 : There is no significant difference between the mean returns of the
stocks that fulfill different principles and the stocks which do not
fulfill the principles
Table 4.12 shows the results of independent sample t-test applied to examine the
differences in the mean returns of the two groups in the holding period of 12 months as
well as 24 months.
109
Table 4.12: Results of t-test Employed on Market Adjusted Returns of the Stocks that Fulfill Graham’s Principles
Compared to the Firms which do not Fulfill Graham’s Principles
Assumption
Differen F-value of Assumptiof F-value of
Differenc of
Principles
ce in Levene’s Equal Mean Levene’s

No. of Mean e in mean Equal
Criteria Std. mean test of Variances t-value returns test of
stocks returns Std. Dev. returns Variances t-value
Dev. returns equality of /No Equal (Annualiz equality of
(Annual) of two /No Equal
of two variance Variances ed) variance
groups Variances
groups Assumed
Assumed
Fulfillment 604 21.218 55.228 9.398 30.068

No equal equal
P1 Non- 21.071 13.553*** 8.146*** 9.8 0.125 6.637***
1305 0.1464 46.27 variance -0.401 29.97 variance
fulfillment
Fulfillment 531 13.208 58.567 5.672 34.07
No equal No equal
P2 Non- 8.858 20.05*** 3.128*** 4.118 17.78*** 2.469**
1378 4.349 46.413 variance 1.553 28.705 variance
fulfillment
Fulfillment 325 17.843 52.821 7.372 29.417
equal equal
P3 Non- 13.297 1.447 4.366*** 5.673 0.432 3.058***
fulfillment
Fulfillment 624 17.198 57.613 9.429 32.3
No equal No equal
P4 Non- 15.415 33.66*** 5.858*** 9.998 10.715*** 6.568***
fulfillment
Fulfillment 91 37.176 66.527 20.72 33.133
No equal equal
P5 Non- 31.882 19.742*** 4.511*** 18.923 1.826 5.857***
fulfillment
Fulfillment 1731 7.314 49.614 3.252 29.709
No equal No equal
P6 Non- 5.369 6.541** 1.235 5.941 10.303*** 2.154**
fulfillment
Fulfillment 853 6.681 51.432 2.874 30.77
equal equal
P7 Non- -0.237 0.018 -0.103 0.316 0.013 0.227
1056 6.919 49.265 variance 2.557 30 variance
fulfillment
Fulfillment 914 9.425 47.635 4.486 28.624
No equal No equal
P8 Non- 5.011 7.022*** 2.189** 3.429 7.528*** 2.481**
fulfillment
P9 Fulfillment 884 7.641 49.167 3.347 29.035
equal No equal
Non- 1.543 2.645 0.669 1.208 7.065*** 0.873
fulfillment
P1 Fulfillment 681 9.394 50.138 4.243 29
equal No equal
Non- 0 4.013 1.713 1.673* 2.401 4.211** 1.69*
fulfillment
Note: *, **, *** denotes p-values significant at 10, 5 and 1 percent level respectively
110
Table 4.12 shows the results of independent sample t-test employed to examine
the difference in the mean market adjusted returns of the stocks that fulfill different
principles compared to the stocks which do not fulfill the said principles. Table 4.12
reveals that the Levene‟s test intended to examine the difference in variance in two
groups; one meeting the principle of earnings yield (P1) and another not meeting the
principle, has significant value of F-statistic. It therefore leads to the rejection of null
hypothesis of equal variance in two groups. Thus, the results of Welch‟s t-test become
applicable in such a situation. Further, in case of 24 months holding of stocks, the
insignificant value of F-statistic leads to the acceptance of null hypothesis of equal
variance in two groups. Thus, the distortions around the mean tend to equalize with the
passage of time (Malkiel, 2012). Table 4.12 further shows that the mean market adjusted
returns of the stocks that fulfill the criterion of earnings yield (21.218% in case of 12
months, 9.398% in case of 24 months holding period) are significantly higher than the
mean market adjusted return of the stocks which do not fulfill this principle (0.146% in
case of 12 months, -0.401% in case of 24 months). Similarly, the differences in the mean
market adjusted return of stocks that meet the principle of price shrinkage (P2), dividend
yield (P3), price to book value (P4), price to net current asset value (P5), debt to net
current asset value (P8) and earnings stability (P10), when compared with the mean
market adjusted returns of the stocks which do not fulfill these principles, turned out to be
significant in both of the holding periods. Also, the significant difference is found
between the mean market adjusted return of the stocks that meet the criterion of having
lesser debt in relation to book value (P6) in comparison with the stocks which do not
meet this criterion in case of 24 months holding period. However, in case of 12 months
holding period, the difference between the two has been insignificant. Furthermore, when
the differences in the mean market adjusted return of the groups classified on the basis of
fulfillment or non- fulfillment of the principle of current asset ratio (P7) and earnings
growth (P9) is examined, the difference is found to be insignificant in both the holding
periods.
111
Hence on the basis of above discussion, we find that the principle of current ratio
being greater than two and a substantial growth in earnings over a period of time, has not
been able to bring out any difference in returns. Thus, any investment decision on the
basis of these two principles alone would not be fruitful in context of Indian stock
market.
It is important to see that all the portfolios are generating significantly greater
returns than the market across the period of study. Thus, Graham‟s stock selection criteria
help an investor to outperform the market. However, the above result covers only the
market factor. It is not necessary that a portfolio outperforming the market is a
dependable portfolio since if in a particular period the market is generating the return of -
15% and a stock meeting the Graham‟s criteria is generating a return of -12%. In such a
situation, the market adjusted return of the stock is 3% [(-12%)-(-15%)]. Thus, in spite of
negative returns generated by a portfolio, the market adjusted returns of the portfolio are
positive. The portfolio no doubt is performing above the market but is not able to deliver
anything in investor‟s basket. Had an investor invested in risk free government securities,
he would have definitely earned something in such a situation. Hence, the capital asset
pricing model (CAPM) helps to analyze the returns and the riskiness of the stocks
considering the risk free rate on government securities and market risk premium.
4.2.3 Analyzing the Risk and Return Relationship of Stocks through Capital Asset
Pricing Model
The basic objective of an investor is to maximize the return obtained from an
investment and minimize the risk involved. The asset pricing model measures the
relationship between the expected return and the risk of the securities. Based on the
CAPM model, a stock‟s expected return is determined by two things: a) the time value of
money and b) the risk premium. The risk free rate represents the time value of money
which implies that the investors are recompensed for investing money in any investment
avenue over a period of time. Additionally, the extra return that investors demand for
taking on risk is called the risk premium i.e., this return depends on the risk measure, and
the market risk premium (Brealey and Myers, 2003). This model determines the expected
112
return of the portfolio on the basis of systematic risk (beta) contained in total variation. In
order to analyze the risk and return relationship of different stocks arrived at, after
meeting different principles, we applied asset pricing model which measures the relative
riskiness of the said portfolio with respect to market portfolio in terms of beta and
examined the presence of abnormal returns, if any, yielded by such stocks through alpha,
popularly known as Jensen alpha, which is estimated by regressing monthly returns of the
portfolio against the market returns during the period of June 1996 to June 2010 by
following time series equation:
Rpt - R ft p p Rmt - R ft e pt
Where, R pt is the return of portfolio p at time t, R ft is the rate of return on a risk-free
asset, p
is the intercept term, Rmt is the rate of return on the market index, p
is the
coefficient loading for the excess return of the market portfolio over the risk-free rate,
and e pt is the error term for portfolio p at time t.
It is important to note that the yield on 91 days Treasury bill has been taken as
proxy for risk free rate and the returns yielded by Sensex has been taken as proxy for
market returns. In order to apply asset pricing model to assess the abnormal returns, the
excess monthly returns (raw returns in excess of risk free rate) of all the stocks meeting
the said criteria are regressed against the excess market returns (market returns in excess
of risk free rate) for the period of 15 years. The intercept term or alpha in the above
equation will determine the presence of abnormal returns obtained through the stocks.
Therefore to assess the abnormal returns generated, if any, we test the following
hypothesis:
H012 : The abnormal return yielded by different principles of Graham’s
stock selection criteria is equal to zero.
Table 4.13 shows the results of the risk- return relationship of stocks assessed
through asset pricing model
113
Table 4.13: Results of Asset Pricing Model Applied on Stocks Meeting Different
Principles of Graham in case of 12 Months Holding Period
Stocks ANOVA
Alpha Beta
meeting Results
R2 DW
different F- t- p- t- p-
p-value Coefficients Coefficients
principles value value value value value
1.490 .303
P1 0.045 1.91 8.332 0.004*** 1.793 .075* 2.887 .004***
(0.831) (0.105)
1.334 .259
P2 0.038 1.85 6.977 0.009*** 1.719 .087* 2.641 .009***
(.776) (.098)
.671 .188
P3 0.027 1.85 4.980 0.026** 1.006 .315 2.231 .026**
(0.667) (.084)
1.487 .235
P4 0.024 1.88 4.391 0.037** 1.672 0.096* 2.095 .037**
(0.889) (0.112)
3.383 .363
P5 0.033 1.941 5.701 0.018** 2.818 .005*** 2.388 .018**
(1.200) (.152)
1.163 .198
P6 0.027 1.888 4.885 0.028** 1.644 .102 2.210 .028**
(.708) (.089)
1.135 .195
P7 0.024 1.911 4.448 0.036** 1.554 .122 2.109 .036**
(.731) (.092)
1.375 .186
P8 0.027 1.886 4.99 0.027** 2.092 .038** 2.234 .027**
(.657) (.083)
1.343 .221
P9 0.035 1.92 6.417 0.012** 1.945 .053* 2.533 .012**
(.690) (.087)
1.338 .213
P10 0.034 1.901 6.249 0.013** 1.985 .049** 2.500 .013**
(.674) (.085)
Capital asset pricing model (CAPM) measures the relation between the expected
return and risk of the securities. From Table 4.13 we notice that the ANOVA, which
measures the goodness of the fit of the model, has F-value significant at 1% level of
significance in case of P1 and P2. The F-value of the ANOVA is significant at 5% level
of significance in case of P3, P4, P5, P6, P7, P8, P9 and P10. Hence, the overall model is
fit. Also, the value of Durbin Watson (DW) in case of all the principles is close to 2
suggesting that there is no autocorrelation in the data.
As per asset pricing model, beta is the only applicable estimate of a stock's risk. It
measures how much a stock is volatile in relation to overall market i.e. how much a
stock‟s market price moves up and down in comparison with how much the stock market
index (Sensex used as the proxy) moves up and down. It can be seen that the value of
beta coefficient is 0.303 in case of P1, 0.259 in case of P2, 0.188 in case of P3, 0.235 in
case of P4, 0.363 in case of P5, 0.198 in case of P6, 0.195 in case of P7, 0.186 in case of
P8, 0.221 in case of P9 and 0.213 in case of P10, when the stocks have been held for the
114
period of 12 months. It shows that 1% increase (decrease) in market portfolio will result
in about 0.303% increase (decrease) in the portfolio of stocks having at least twice the
yield on AAA corporate bonds, 0.259% increase (decrease) in case of stocks having
lesser current PE than the 40% of the highest PE the stocks had over past 5 years, 0.188%
increase (decrease) in the portfolio of stocks having minimum dividend yield of at least
two-third the AAA bond yield, 0.235% increase (decrease) in case of stocks having
current price lesser than two- third of the tangible book value per share and about 0.363%
increase (decrease) in case of stocks having current price lesser than the net current asset
value per share of the stock. Similarly, 1% proportionate change in market portfolio is
followed by about 0.198% change in the stocks having lesser debt than the book value,
0.195% change in case of portfolio having current ratio greater than 2, 0.186% of up and
down in case of stocks having total debt lesser than twice the net current assets, followed
by a change of about 0.221% in case of stocks showing growth in their earnings over a
period of time and 0.213% change in case of stocks having stability in their per share
earnings over a period of time. Also, the beta is significant at 1% level of significance in
case of P1 and P2. Beta is significant at 5% level of significance in case of P3, P4, P5,
P6, P7, P8, P9, and P10. Moreover, beta is the significant factor explaining the variation
in portfolio‟s returns. Also, the beta value lesser than one, suggests that the portfolio
under study is lesser volatile or risky than the market. Therefore, the stocks selected
through Graham‟s stock selection criteria do not rise and fall as briskly as the overall
market does.
The capital asset pricing model helps us to determine the expected returns of the
portfolio by adequately pricing the systematic risk factor i.e. beta in the model and then
compares the actual returns with the expected return of the portfolio to determine the
presence of abnormal returns (alpha). The Jensen alpha as discussed above explains the
difference between the portfolio‟s actual return and expected return, which is significant
at 1% level of significance in case of P5 i.e. the stocks having lesser current price than
two-third of its net current asset value per share, implying that the stocks trading at lesser
115
price than their liquidation value offer good margin of safety to the investors. Such stocks
therefore take no time to revert and the investors can really expect the instantaneous
increase in the price of such stocks. The investors hence can enjoy abnormal returns by
investing in such stocks. Further, the value of the alpha has been significant at 5% level
of significance in case of P8 suggesting that the investors do not want to rely much on
highly levered company. They want the company to have adequate resources to cover
their debt. Also, the alpha is significant at 5% level of significance in case of P10,
suggests that stability in the growth of earnings can provide important clue to the
financial soundness and the good reputation of the company. Thus, the investors
considering the principle of investing in company having current asset base to cover it‟s
all liabilities as well as generating stable earnings over a period of time, can help the
investors to reap excess returns in market in case of 12 months holding period of the
stocks. In case of P1, P2, P4 and P9, the excess returns are statistically significant only at
10% level of significance.
In case of principle of high dividend yield (P3), principle of total debt being lesser
than the total book value (P6) and the principle of high current ratio (P7), the value of
alpha is positive but statistically insignificant. It shows that stocks selected and held for
the period of one year do not provide any excess returns to investors. Also, the R-square
that gives the proportion of variance explained by the regression model or the market
factor is very small i.e. 4.5% in case of P1, 3.8% in case of P2, 2.7% in case of P3, 2.4%
in case of P4, 3.3% in case of P5, 2.7% in case of P6, 2.4% in case of P7, 2.7% in case of
P8, 3.5% in case of P9 and 3.4% in case of P10. Thus it implies that the market factor
plays a small role in explaining the variation in overall returns.
Table 4.14 shows the results of asset pricing model when the stocks selected
through the Graham‟s principles have been held for the period of 24 months.
116
Principles of Graham in Case of 24 Months Holding Period
Stocks ANOVA Results Alpha Beta

meeting 2
R DW F- t- t-
different P-value Coefficients p-value Coefficients p-value
value value value
principles
1.458 .301
P1 0.041 1.941 15.368 0.00*** 2.421 .016** 3.920 .000***
(.602) (.077)
1.285 .266
P2 0.037 1.887 13.852 0.00*** 2.290 .023** 3.722 .000***
(.561) (.072)
.882 .159
P3 0.011 1.845 4.312 0.038** 1.468 .142 2.076 .038**
(.601) (.076)
1.508 .490
P4 0.033 1.905 12.546 0.000*** 2.412 .016** 3.540 .000***
(.625) (.138)
2.912 .277
P5 0.022 1.861 7.414 0.007*** 3.652 .000*** 2.723 .007***
(.797) (.102)
1.156 .205
P6 0.028 1.912 10.385 0.001*** 2.321 .021** 3.223 .001***
(.498) (.063)
1.106 .206
P7 0.026 1.953 9.669 0.002*** 2.131 .034** 3.110 .002***
(.519) (.066)
1.259 .186
P8 0.027 1.906 10.116 0.002*** 2.749 .006*** 3.181 .002***
(.458) (.058)
1.223 .231
P9 0.037 1.95 13.616 0.00*** 2.492 .013** 3.690 .000***
(.491) (.063)
1.176 .228
P10 0.038 1.933 14.022 0.00*** 2.459 .014** 3.745 .000***
(.478) (.061)
From Table 4.14, we notice that when the holding period of the stocks has been
extended from 12 months to 24 months, the P-value of the ANOVA is found to be
significant at 1% level of significance in case of all principles except the principle of
dividend yield, where it is significant at 5% level of significance. Thus, the overall model
is fit. Also, the value of Durbin Watson in case of all the principles is close to 2
suggesting that there is no problem of autocorrelation in the data. The value of beta
coefficient in case of P1 is 0.301 showing that 1% change in market index would result in
about 0.301% change in stocks having twice the yield on AAA bonds. Similarly 1%
change in market portfolio will results in the change of about 0.266% in case of P2,
0.159% in case of P3, 0.490% in case of P4 and 0.277% in case of P5. Also, 1% up and
down in Sensex would result in 0.205% increase or decrease in case of P6, 0.206% in
case of P7, 0.186% in case of P8, 0.231% in case of P9 and 0.228% in case of P10. It
117
shows that even when we extend the holding time of the portfolio from 12 months to 24
months, the stocks selected through Graham‟s stock selection criteria remain lesser risky
and lesser volatile than the market.
The R-square which measures the proportion of variance explained by the
independent variable i.e. market risk factor „beta‟ in case of 24 months holding period is
4.1% in case of P1, 3.7% in case of P2, 1.1% in case of P3, 3.3% in case of P4, 2.2% in
case of P5, 2.8% in case of P6, 2.6% in case of P7, 2.7% in case of P8, 3.7% in case of P9
and 3.8% in case of P10. The results therefore substantiate the small explanatory power
of the model.
Further, the value of intercept or alpha again is significant at 1% level of
significance in case of P5 showing that the stocks that have lesser market price than two-
third of their net current asset value continue to yield returns in excess of market risk
factor and risk free government security, even after extending the holding period to two
years. Also, the stocks selected on the principle of total debt being lesser than twice the
net current asset value are generating abnormal returns significant at 1% level of
significance followed by 5% level in case of P1, P2, P4, P6, P7, P9 and P10. However in
case of P3 (dividend yield), the alpha has not been significant showing that the investors
cannot reap any excess returns if they invest solely on the basis of principle of dividend
yield being at least two- third the AAA bond yield.
4.3 TESTING THE PERFORMANCE OF DIFFERENT COMBINATIONS OF

GRAHAM’S STOCK SELECTION CRITERIA
The above analysis of Graham‟s stock selection criteria considered every
principle individually in analyzing their potency in providing excess return to investors.
However, Graham recommended that the stock that meets each and every principle of
stock selection should be bought by the investor. But the stock selection rules of
Benjamin Graham have been considered quite stringent by the researchers who consider
that if all the rules of Graham are applied simultaneously on the stocks, not even a single
stock will qualify for the portfolio (Klerck and Maritz, 1997). Also, the argument has
118
been raised that the time when these rules were given, cheap stocks were available in
market but it is quite difficult to find such stocks meeting all the rules in the present
environment. Thus, in reality, much of Graham‟s work would be difficult to apply to the
modern stock market world 5 . Therefore, In order to examine the applicability of
Graham‟s principles in present scenario, different researchers have formed the reward-
risk combinations of the stocks in order to examine the performance of overall criteria
(Oppenheimer, 1984; Klerk and Maritz, 1997).
Benjamin Graham‟s ten principles of stock selection construct a formidable
risk/reward barrier. The first five principles i.e. P1, P2 P3, P4 and P5 point towards
potential reward by pinpointing a low price in relation to such key operating results as
earnings and are sensitive to price and earnings changes. The second five principles i.e.
P6, P7, P, P9 and P10 measure risk by measuring financial soundness and stability of
earnings and does not change rapidly with the changes in price and earnings (Phalon,
2001). Thus, an attempt has been made to examine the profitability of different risk-
reward combinations of Graham‟s stock selection criteria.
In order to screen stocks on the basis of risk- reward combination, the stocks are
screened meeting one principle of reward measures and another principle conforming risk
measures as advocated in Oppenheimer (1984), Klerk and Maritz (1997). It therefore
results in the total of 25 risk- reward combinations i.e. P1-P6, P1-P7, P1-P8, P1-P9, P1-
P10, P2-P6, P2-P7, P2-P8, P2-P9, P2-P10, P3-P6, P3-P7, P3-P8, P3-P9, P3-P10, P4-P6,
P4- P7, P4-P8, P4-P9, P4-P10, P5-P6, P5-P7, P5-P8, P5- P9 and P5-P10. Further, each
year, the stocks have been screened on the basis of clearance of these risk- reward
combinations. Table 4.15 presents the distribution of stocks amongst different risk-
reward combinations across the period of study.
5
http://www.stockexchangesecrets.com/value-investing-rules-of-ben-graham.html
119
Table 4.15: Distribution of Stocks Amongst Different Risk- Reward Combinations Across the Period of Study (1996-2010)
Year 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Total No. of stocks
Combinations
P1-P6 2 14 20 16 29 45 28 47 63 39 20 20 25 103 52 523
P1-P7 0 7 13 7 15 33 18 25 20 18 12 11 18 57 36 290
P1-P8 1 2 4 6 12 17 8 21 28 14 11 9 12 51 20 216
P1-P9 2 6 12 9 14 20 11 26 36 23 16 12 20 58 45 310
P1-P10 3 6 11 9 10 16 9 16 22 15 8 9 14 43 26 217
P2-P6 32 28 39 38 36 42 24 34 35 16 10 22 39 42 38 475
P2-P7 13 14 22 16 18 30 13 17 13 6 6 10 24 26 17 245
P2-P8 10 7 11 12 14 20 12 21 20 10 5 11 21 19 13 206
P2-P9 22 15 28 22 19 22 14 22 18 8 9 16 23 25 32 295
P2-P10 12 12 26 19 14 16 9 11 9 3 2 9 23 18 18 201
P3-P6 2 12 17 17 20 18 19 34 44 17 12 6 6 58 12 294
P3-P7 1 7 9 9 12 12 9 15 11 7 7 4 3 28 5 139
P3-P8 0 1 0 5 4 3 2 17 23 10 9 4 4 31 4 117
P3-P9 0 4 8 8 8 4 4 16 19 5 8 4 2 28 7 125
P3-P10 0 4 7 8 9 6 4 9 14 5 3 3 2 20 4 98
P4-P6 13 22 25 36 41 51 39 41 39 27 28 22 40 56 50 530
P4-P7 4 12 16 16 21 35 19 19 11 6 9 15 26 31 28 268
P4-P8 4 3 4 13 16 19 10 14 18 13 14 12 20 25 23 208
P4-P9 7 7 14 14 15 17 8 12 17 12 15 13 20 27 33 231
P4-P10 6 7 14 15 13 16 7 5 11 4 6 7 14 16 14 155
P5-P6 0 1 1 5 10 9 5 5 8 6 5 5 10 11 9 90
P5-P7 0 0 0 2 4 4 0 1 1 1 2 2 4 3 3 27
P5-P8 0 1 1 5 10 8 5 5 8 6 6 5 10 11 9 90
P5-P9 0 0 1 0 4 1 0 2 5 3 4 4 4 5 6 39
P5-P10 0 1 1 1 2 2 1 2 3 1 2 2 5 3 1 27
120
As evident from Table 4.15 that when the stocks meeting the principle of earnings
yield (P1) along with debt equity ratio (P6) are screened, the number of stocks meeting
both the criteria over the period of 15 years is 523. Similarly, the principle of earnings
yield (P1) when clubbed with the principle of total debt lesser than twice the net current
assets (P8) yields 216 stocks in the basket across the period of 15 years. It is important to
note from Table 4.15 that no stock could be screened in year 1996 on the basis of
combinations like P1-P7, P3-P8, P3-P9, P3-P10, P5-P6, P5-P7, P5-P8, P5-P9 and P5-
P10. Also in 1997, no stock could be obtained in case of P5-P7 and P5-P9. Similarly, the
number of stocks in rest of the combinations can be elicited from the above Table. Thus,
the distribution of stocks shows that there are no or very few stocks in certain years (0, 1,
2, 3 etc.). Hence, the stocks on basis of different risk -reward combinations screened
across the period of the study are analyzed to determine the statistical significance of
excess returns yielded, if any.
4.3.1 Analyzing the Market Adjusted Performance of the Combinations

The stocks are screened on the basis of different risk- reward combinations across
the period of the study and their market adjusted returns are computed for the purpose of
analysis. The following hypothesis is thus examined:
H013 : The market adjusted return of different risk-reward combinations of
Graham’s criteria is equal to zero.
Table 4.16 shows the performance of the risk- reward combinations as under:
Meeting Different Risk-Reward Combinations
No. of
Combinations Mean Std. t- p- Mean Std. t- p-
stocks
(Annual) Dev. Value Value (Annualized) Dev. Value Value
P1-P6 22.4416 10.5953
523 55.41852 9.261 .000*** 28.98648 8.359 .000***
(2.42328) (1.26749)
P1-P7 21.9056 10.5885
290 54.48823 6.846 .000*** 29.25896 6.163 .000***
(3.19966) (1.71815)
P1-P8 26.6711 11.9855
216 52.80235 7.424 .000*** 29.07651 6.058 .000***
(3.59274) (1.97841)
P1-P9 17.2620 8.2860
310 52.59113 5.779 .000*** 28.50000 5.119 .000***
(2.98698) (1.61869)
P1-P10 21.2181 9.3983
217 55.22851 9.442 .000*** 30.06865 7.682 .000***
(2.24722) (1.22348)
P2-P6 14.2890 6.3361
475 58.94545 5.283 .000*** 33.93110 4.070 .000***
(2.70460) (1.55687)
121
P2-P7 19.6662 8.7840

245 58.65261 5.259 .000*** 34.19287 4.029 .000***
(3.73955) (2.18006)
P2-P8 16.9796 7.6060
206 55.18299 4.416 .000*** 33.68935 3.240 .001***
(3.84478) (2.34725)
P2-P9 11.1836 4.2718
295 56.55364 3.396 .001*** 33.86998 2.166 .031**
(3.29268) (1.97199)
P2-P10 16.5970 7.2912
201 58.87853 3.996 .000*** 32.31713 3.199 .002***
(4.15297) (2.27947)
P3-P6 18.8248 7.7292
294 53.21403 6.066 .000*** 29.10331 4.554 .000***
(3.10351) (1.69734)
P3-P7 26.4223 12.1238
139 55.26336 5.637 .000*** 27.55172 5.188 .000***
(4.68738) (2.33691)
P3-P8 11.9844 4.0333
117 47.17867 2.748 .007*** 27.43300 1.590 .114
(4.36167) (2.53618)
P3-P9 17.8798 6.7541
125 54.99203 3.635 .000*** 27.44263 2.752 .007***
(4.91864) (2.45454)
P3-P10 21.4970 12.2829
98 60.35280 3.526 .001*** 27.04996 4.495 .000***
(6.09655) (2.73246)
P4-P6 18.6623 10.8739
530 59.01565 7.280 .000*** 32.13789 7.789 .000***
(2.56348) (1.39598)
P4-P7 19.1810 11.2682
268 53.72375 5.845 .000*** 29.67842 6.216 .000***
(3.28170) (1.81290)
P4-P8 21.0432 11.1984
208 58.63463 5.176 .000*** 33.63897 4.801 .000***
(4.06558) (2.33244)
P4-P9 16.6331 8.0723
231 58.74497 4.303 .000*** 32.31579 3.797 .000***
(3.86513) (2.12622)
P4-P10 23.0387 12.5463
155 62.04946 4.623 .000*** 31.77491 4.916 .000***
(4.98393) (2.55222)
P5-P6 38.0959 20.6706
90 66.31640 5.450 .000*** 33.31595 5.886 .000***
(6.99036) (3.51181)
P5-P7 21.5160 10.4017
27 73.74407 1.516 .142 36.37312 1.486 .149
(14.1925) (7.00001)
P5-P8 36.7988 20.5811
90 66.80224 5.226 .000*** 33.29249 5.865 .000***
(7.04157) (3.50934)
P5-P9 26.2817 15.4270
39 59.23816 2.771 .009*** 31.77063 3.032 .004***
(9.48570) (5.08737)
P5-P10 25.6299 15.1009
27 50.58938 2.633 .014** 24.50788 3.202 .004***
(9.73593) (4.71654)
Note:
Table 4.16 shows the performance of stocks selected on the basis of different risk
reward combinations. It is important to see that all the risk-reward combinations have
yielded higher return than the market across the period of study. Further, when the one
sample t-test has been used to examine the performance of the said combinations, we find
that mean returns have been significantly higher than the market (significant at 1% level
of significance) in case of all the combinations except the combination of price to net
122
current asset ratio (P5) with the earnings stability (P10), where the returns have been
significant at 5% level of significance. It is important to notice that when the principle of
price being lesser than two- third of the net current asset value of the stock (P5) is
clubbed with the principle of current ratio being greater than two (P7), the mean market
adjusted return of 21.51% across the period of study has been insignificant. It could be
due to the fact that larger current ratio could imply excessive unutilized inventory which
leads to low profitability [see, for example, Soenen (1993), Shin and Soenen (1998),
Wang (2002), Deloof (2003), Eljelly (2004), Lazaridis and Tryfonidis (2006), Padachi
(2006), Raheman and Nasr (2007), Garcia-Teruel and Martinez-Solano (2007), Ganesan
(2007), Azhar and Saad (2010), Gill et. al (2010), Dong and Su (2010), Danuletiu (2010),
Izadinia and Taki (2010), Hayajneh and Yassine (2011), Vahid et. al (2012)]. It may be
due to the reason that a high current ratio implies heavy investment in current assets
which may mean under utilization of resources or more liquidity than required which
leads to lesser profitability (Chary, et. al, 2011). The principle of buying the stocks when
current price is lesser than two-third of the net current asset value has yielded the highest
as well as significantly positive market adjusted returns along with abnormal returns
when examined individually (refer Table 4.5). However, when this principle has been
clubbed with the principle of current ratio, to screen stocks, the returns so attained
through those stocks has been insignificant. This could imply that the stocks are trading
at such a low price due to the fact that the stocks have high level of inventory which is
not utilized efficiently. Thus, the earning power of such firms is low. High level of
inventory also leads to high holding cost for the firm and the funds that could have been
used otherwise in research and development, marketing etc. gets blocked in inventory.
However, rest all the combinations can be used safely to attain significantly higher
returns than the market in case of 12 months holding period.
Further analysis shows that even after extending the holding period to 24 months,
the different risk- reward combinations continue to provide higher returns than the
market. Also, the t-test employed to judge the significance of market adjusted returns
shows that keeping aside the combination of P3-P8, P5-P7, rest all the combinations are
generating significantly positive market adjusted returns to investors.
123
It is imperative to see that in case of combination of P5-P7, the returns remain

insignificant even when the holding period of the stocks has been extended from 12
months to 24 months. Thus, the insignificant returns yielded by the combination implies
the lack of efficient utilization of the blocked funds. Along with it, the principle of
dividend yield (P3) when clubbed with the principle of total debt being lesser than twice
the net current assets (P8) also yields insignificant returns to the investors. It could be due
to the reason that total debt being lesser than twice the net current assets means very low
debt owed by the company. Such companies are using owned funds rather than borrowed
funds to conduct its operations. Along with it, such firms have high dividend yield and
hence lesser retained funds. It, therefore, implies that such firms have lesser growth
opportunities so they neither opt for borrowed funds nor go for retained funds to encash
future opportunities. Such companies would have lesser possibility of growing in future.
Also, the financing from borrowed funds has become an essential component of a firm‟s
total finance these days. Moreover, when benefit and cost of debt are simultaneously
considered, the leverage is positively related to the firm value before reaching firm‟s
optimal capital structure (Cheng and Tzeng, 2011). Thus, one can ignore this
combination while using Graham‟s criteria for investment in stocks.
4.2.3 Analyzing the Presence of Abnormal Return Yielded by Combinations

The above performance of the different risk- reward combinations of the ten rules
was based on market risk factor only. The asset pricing model employed on same
combination of stocks further help us to know if any excess returns generated to investors
after capturing the time effect of the money along with the market risk premium. Thus,
we purport to examine the following hypothesis:
H014 : The abnormal return yielded by different risk-reward combinations
of Graham’s stock selection criteria is equal to zero.
Table 4.17 reports the results of testing of above hypothesis.
124
Combinations of Principles of Graham in case of 12 Months Holding Period
ANOVA Results Alpha Beta
Combinat
R2 DW F- Coefficie t- p- Coefficie t-
ions p-value p-value
value nts value value nts value
0.004** 1.727 .307
P1-P6 0.046 1.861 8.504 2.073 .040** 2.916 .004***
* (.833) (.105)
0.004** 2.079 .343
P1-P7 0.049 1.983 8.551 2.240 .026** 2.924 .004***
* (.928) (.117)
0.004** 2.041 .324
P1-P8 0.045 1.946 8.335 2.300 .023** 2.887 .004***
* (.887) (.112)
0.007** 1.517 .284
P1-P9 0.04 2.002 7.392 1.834 .068* 2.719 .007***
* (.828) (.105)
1.718 .252
P1-P10 0.03 1.947 5.583 0.019** 2.038 .043** 2.363 .019**
(.843) (.107)
0.009** 1.442 .259
P2-P6 0.038 1.837 7.064 1.874 .063* 2.658 .009***
* (.770) (.097)
1.629 .250
P2-P7 0.032 1.949 5.955 0.016** 2.011 .046** 2.440 .016**
(.810) (.102)
0.005** 1.582 .277
P2-P8 0.044 1.841 8.165 2.064 .041** 2.857 .005***
* (.767) (.097)
1.347 .227
P2-P9 0.03 1.913 5.59 0.019** 1.774 .078* 2.364 .019**
(.759) (.096)
1.791 .194
P2-P10 0.022 1.821 4.038 0.046** 2.350 .020** 2.009 .046**
(.762) (.096)
.860 .281
P3-P6 0.036 1.877 6.838 .009*** 1.011 .313 2.614 .009***
(.851) (.107)
0.813 0.198
P3-P7 0.027 1.91 5.830 0.016** 1.149 0.252 2.414 0.016**
(0.707) (0.082)
P3-P8 - - - - - - - - - -
.952 .213
P3-P9 0.019 1.99 3.244 0.073* 1.019 .309 1.801 .073*
(.934) (.118)
1.996 0.263
P3-P10 0.034 1.771 5.613 0.018** 2.384 .018** 2.369 0.018**
(0.837) (.111)
1.594 .223
P4-P6 0.022 1.869 4.059 0.045** 1.822 .070* 2.015 .045**
(.875) (.111)
1.564 .314
P4-P7 0.042 2.051 7.941 .005** 1.773 .077* 2.818 .005***
(.882) (.111)
1.629 .271
P4-P8 0.031 1.824 5.609 0.019** 1.797 .074* 2.368 .019**
(.906) (.115)
1.709 .250
P4-P9 0.027 1.912 4.926 0.028** 1.917 .057* 2.219 .028**
(.891) (.113)
2.026 .245
P4-P10 0.021 1.918 3.811 0.052* 2.043 .042** 1.952 .052*
(.991) (.125)
2.910 .008** .341
P5-P6 .0353 1.744 6.085 .0146** 2.661 2.466 .0146**
(1.093) * (.138)
P5-P7 - - - - - - - - - -
3.421 .005** .359
P5-P8 0.032 1.929 5.53 0.02** 2.833 2.352 .020**
(1.208) * (.153)
P5-P9 - - - - - - - - - -
1.996 .263
P5-P10 0.039 2.001 3.910 .049** 2.455 .015** 1.977 .049**
(0.813) (.133)
Note:
2. Standard error of the coefficients have been reported in parenthesis
3. The stocks pertaining to combinations; P3-P8, P5-P7, P5-P9 have been zero in certain years,
due to which the time series data for those years goes missing. Hence, the time series CAPM
model could not be applied on these combinations.
125
Table 4.17 shows that the asset pricing model when applied on different risk-
reward combination of stocks reveals that the P-value of the ANOVA has been
significant in case of all the combinations showing the goodness of fit of the model.
Further, the value of beta is lesser than 1 in all the combinations showing that all the risk-
reward combinations suggested by Graham have lesser volatility than the market
portfolio and therefore lesser risky than the market.
Out of all the combinations tested, maximum number of combinations has been
generating significantly excess returns (determined through alpha) to the investors. It is
important to note that the alpha is significant at 1% level of significance in case of stocks
having market price lesser than two-third of their net current asset value (P5) along with
total debt lesser than their book value (P6). The same thing holds when the stocks having
lesser price than their net current asset value (P5) are screened with the condition that the
total debt is lesser than twice the net current asset value (P8). The excess return
determined in case of both the combinations suggest that the investors can safely invest in
securities trading below their liquidation value, given that such companies have very low
level of outside liabilities. Such securities give investors the opportunity to unlock the
intrinsic value in the period of one year holding of such securities. Similarly, the
combinations such as P1-P6, P1-P7, P1-P8, P1-P10, P2-P7, P2-P8, P2-P10, P3-P10, P4-
P10 and P5-P10 are yielding excess returns significant at 5% level of significance.
Further, the combinations such as P1-P9, P2-P6, P2-P9, P4-P6, P4-P7, P4-P8, P4-P9, and
are generating excess returns significant at 10% level of significance only, when held for
the period of 12 months.
Table 4.17 further shows that the significant excess returns cannot be determined
when the stocks having dividend yield of at least two- third the bond yield (P3) are
screened along with the stocks having lesser debt than the book value (P6). It could be
due to the reason that total debt being lesser than the book value means very low debt
owed by the company. Such companies are hence using owned funds rather than
borrowed funds to carry out its operations. Along with it, such firms have high dividend
yield or higher payout ratio. Larger dividends result in less retained earnings, which are
126
necessary for financing growth and modernization of firm, which may in turn hamper
growth rate in earnings and share price (Shukla, 2011). Such companies would have
lesser possibility of growing in future and thus, excess returns could not be retrieved by
investing in stocks having combination of high dividend yield and very low level of debt.
Along with it, the principle of high dividend yield (P3) when clubbed with the principle
of growth in earnings (P9) is also not able to yield significant abnormal returns in Indian
stock market.
Along with it, when the stocks having dividend yield of at least two-third the
bond yield (P3) are screened along with the stocks having current ratio greater than two
(P7), significant excess returns cannot be assessed in case of one year holding period of
the stocks. This combination could also mean lesser level of retained funds with the
company and huge investment blocked in current assets. Thus, the stocks selected on the
basis of combination (P3-P7) will not yield any excess returns to investors, if held for the
period of 12 months.
Table 4.18 reports the results of asset pricing model applied on the same stocks
when the holding period has been extended from 12 months to 24 months.
Combinations of Principles of Graham in case of 24 Months Holding
Period
Combinations R2 DW f- f-
p-value Coefficients p-value
value value
1.721 .291
P1-P6 0.039 1.952 14.628 0.000*** 2.881 .004*** 3.825 .000***
(.597) (.076)
1.905 .282
P1-P7 0.033 1.902 11.463 0.001*** 2.915 .004*** 3.386 .001***
(.654) (.083)
1.698 .297
P1-P8 0.037 2 13.825 0.000*** 2.706 .007*** 3.718 .000***
(.627) (.080)
1.593 .288
P1-P9 0.038 2.019 14.079 0.000*** 2.642 .009*** 3.752 .000***
(.603) (.077)
1.750 .258
P1-P10 0.031 1.974 11.394 0.001*** 2.921 .004*** 3.375 .001***
(.599) (.076)
1.352 .261
P2-P6 0.037 1.874 13.789 0.000*** 2.447 .015** 3.713 .000***
(.552) (.070)
1.435 .237
P2-P7 0.026 1.992 9.578 0.002*** 2.391 .017** 3.095 .002***
(.600) (.077)
1.458 .262
P2-P8 0.039 1.893 14.545 0.000*** 2.705 .007*** 3.814 .000***
(.539) (.069)
1.301 .244
P2-P9 0.032 1.914 11.929 0.001*** 2.351 .019** 3.454 .001***
(.554) (.071)
1.398 .229
P2-P10 0.029 1.824 10.786 0.001*** 2.558 .011** 3.284 .001***
(.547) (.070)
127
.585 .165
P3-P6 .0124 1.876 4.505 .0344** .955 .339 2.122 .034**
(.612) (.078)
.684 .219
P3-P7 .0178 1.831 6.518 .011** 1.015 .3107 2.553 .011**
(.674) (.085)
P3-P8 - - - - - - - - - -
.505 .224
P3-P9 0.021 1.933 7.341 .007*** .778 .431 2.709 .007***
(.649) (.082)
1.883 .191
P3-P10 0.015 1.958 5.363 .0211** 2.910 .003*** 2.315 .0211**
(.647) (.082)
1.638 .233
P4-P6 0.024 1.898 8.837 0.003*** 2.665 .008*** 2.973 .003***
(.614) (.078)
1.473 .225
P4-P7 0.021 1.957 7.655 0.006*** 2.312 .021** 2.767 .006***
(.637) (.081)
1.640 .251
P4-P8 0.026 1.86 9.591 0.002*** 2.575 .010** 3.097 .002***
(.637) (.081)
1.624 .269
P4-P9 0.03 1.925 11.21 0.001*** 2.574 .010** 3.348 .001***
(.631) (.080)
1.815 .288
P4-P10 0.029 1.96 10.876 0.001*** 2.650 .008*** 3.298 .001***
(.685) (.087)
2.567 .241
P5-P6 0.020 1.867 7.017 0.008*** 3.583 .000*** 2.649 .008***
(.716) (.091)
P5-P7 - - - - - - - - - -
2.937 .274
P5-P8 0.021 1.853 7.195 0.008*** 3.669 .000*** 2.682 .008***
(.801) (.102)
P5-P9 - - - - - - - - - -
2.000 .230
P5-P10 0.012 2.025 4.22 0.041** 2.281 .023** 2.054 .041**
(.877) (.112)
Note:
2. Standard error of the coefficients have been reported in parenthesis
3. The stocks pertaining to combinations; P3-P8, P5-P7, P5-P9 have been zero in certain years,
due to which the time series data for those years goes missing. Hence, the time series CAPM
model could not be applied on these combinations
Table 4.18 shows that the p-value of the ANOVA has been significant in case of
all the combinations showing the goodness of the fit of the model. Also, the value of
Durbin Watson statistic is close to 2 in all the cases showing no problem of
autocorrelation in the data. Also, the value of beta is lesser than 1 in all the combinations
showing that all the risk- reward combinations suggested by Graham have lesser
volatility than the market portfolio and therefore lesser risky than the market.
It is important to note from Table 4.18 that when the holding period of the stocks
has been extended from 12 months to 24 months, the portfolios which were initially
generating excess returns significant at 10% level of significance (P2-P6, P2-P9, P4-P7,
P4-P8 and P4-P9), have started generating excess returns significant at 5% level of
significance. Also, the portfolios which were initially generating abnormal returns
significant at 5% level of significance (P1-P6, P1-P7, P1-P8, P1-P10, P2-P8, P3-P10, and
P4-P10) have now become significant at 1% level of significance. The critics of efficient
128
market hypothesis state that there exist certain anomalies in the market which the
investors can use for the profitable trading in the market. Till the time the market is not
aware of the anomaly, those who invest on the basis of said anomaly can reap abnormal
returns. “However, the anomalies are inherently self-destructive because public
knowledge of the anomaly will increase the funds allocated to the specific anomaly and
reduce risk-adjusted abnormal returns” (Malkiel, 2005 cited in Guenester, 2009). Also,
Yen et al. (2004) found that value premium is concentrated up to 2 years of portfolio
formation and not afterwards. Thus, an investor can choose the said risk- reward
combination of stocks to exploit the abnormal returns generated through the anomaly of
greater intrinsic value than the market price, especially when the portfolio has been held
for the period of 24 months.
It can also be elicited from Table 4.18 that the stocks selected on the basis of
combinations such as P3-P6 and P3-P7, continue to generate insignificant excess returns
even after extending the holding period of such stocks from 12 months to 24 months.
Thus, an investor can ignore these risk-reward combinations while taking a decision to
invest in Indian stock market. Along with these combinations, the combination of stocks
having high dividend yield (P3) and growth in per share earnings (P9), is also not able to
yield abnormal returns in Indian stock market. Also, Adelegan (2000) suggested that as
growth in earnings increases, dividend yield is higher for low growth firms than for a
high growth firm which is consistent with a high opportunity cost of dividend payout for
high growth firms. Thus, the firms that have high growth in earnings tend to have lesser
payout ratios as such firms require retained funds for financing growth and
modernization. Also, Walter (1963) proposed a model on dividend policy stating that the
company should retain its earnings if the return on investment (r) is greater than the cost
of capital (k). However, if the internal rate of return is lesser than the cost of capital or
capitalization rate, then the earnings should be distributed to shareholders in the form of
dividends. If a firm has adequate profitable investment opportunities, it will be able to
earn more than what the investors expect so that r > k. Such firms are called “growth
firms”, which should plough back the entire earnings within the firm (Shukla, 2011).
Hence, the combination of high earnings growth and high dividend yield, being contrary
to each other, could not yield excess returns in Indian stock market.
129
4.4 REASONS FOR THE EXISTENCE OF VALUE PREMIUM

From the above discussion it is clearly evident that there exists value premium in
Indian stock market. However, the reason behind the existence of the value premium is
debatable one. Different schools of thought have emerged regarding the existence of the
value premium broadly classified into 2: risk based explanation and behavioral
explanation.
The risk based school of thought suggests that the value strategies harvest higher
returns due the fact that the stock selected on the basis of these strategies are riskier than
rest of the market. The value premium is a rational phenomenon, which is priced in
equilibrium, and the value stocks should generate higher average returns as a
compensation for the systematic risk borne by the investor (Ball, 1978; Berk, 1995; Fama
and French, 1998). Thus, the risk-based explanations defend the market efficiency and
suggest that the abnormal profits of the investment strategies can be captured by asset
pricing models and that they do not arise due to the investors‟ irrational behavior
(Hamalainen, 2007).
The risk of a portfolio is measured through its beta coefficient. It is the measure of
the volatility or systematic risk of a stock or portfolio of stocks in comparison to market
as a whole. The market (Sensex is used as proxy) has the beta value of 1. The beta value
of a portfolio lower than 1 suggests that the given portfolio is lesser volatile than the
market and the value greater than 1 suggests that the portfolio has larger volatility than
the market. It is further important to note that all the ten principles of Graham analyzed
individually have lesser beta value than 1 (refer Table 4.1-4.10). Also, the risk- reward
combinations analyzed also showed lesser volatility than the market (see Table 4.17,
4.18). Thus it helps us to safely infer that the stocks selected on the basis of the Graham‟s
stock selection criteria are not risky.
Another school of thought aimed at providing explanation to the existence of
value premium is the behavioral or cognitive biases of individuals which would have
generated value premium to contrarians. The behavioral school of thought suggests that
there are different kinds of biases inherent in the behavior of investors e.g. the investors
generally believe that they are better decision makers than they in fact are. Moreover,
they look for the information that confirms their beliefs. Such self deception leads to
130
choice errors. Individuals become too confident that their decisions are correct and put
too much value on their previous beliefs (Baker and Nofsinger, 2002). Kahneman and
Twersky (1982) first stated that the analysis of human judgment shows that many biases
of intuition stem from the tendency to give little or no weight to certain kind of
information i.e. the forecasters overweigh the more recent information relative to older
data. Different kind of biases originate from this tendency of investors, for example, De
Bondt and Thaler (1985) hypothesized that excess returns on low P/E stocks could be
attributable to overreaction hypothesis. They argued that the investors overreact to both
good and bad news so that the stocks that have risen over last 3 years because of good
news are priced too high and similarly the stocks that have fallen over recent years due to
bad news are priced too low. Consistent with the predictions of overreaction hypothesis,
portfolio of prior losers are found to outperform the prior winners. This is because of the
reason that poor relative performance of winner portfolios are correction of the
mispricing happened due to overreaction. Similarly, Barberis et al. (1998) found that over
longer horizon of perhaps 3 to 5 years, security prices overreact to consistent pattern of
news pointing in same direction i.e. the securities that have had a long record of good
news tend to become over priced and have low average returns afterwards and the
securities that are less popular in market due to bad or no news becomes underpriced and
have high returns afterwards.
Another bias importunate in the behavior of individuals is the familiarity bias.
The investors sometimes put too much emphasis on familiar stocks. Since those stocks
are familiar, investors tend to believe that they are lesser risky than other companies or
even safer than a diversified portfolio (Baker and Nofsinger, 2002). Moreover, the
representativeness bias i.e. the tendency of the investors to buy the stocks that represent
desirable qualities such as significant growth in earnings, sales or assets, also affects the
pricing of securities. Buying the stock of the firm with a history of consistent earnings
growth leads to lower risk adjusted returns in subsequent period (Solt and Statman, 1989)
due to the fact that very few companies can sustain the high level of growth achieved in
the past. Thus, on average, the more the growth opportunities of a company at a given
time, lower the risk adjusted return that its stock provides to the shareholders. On the
contrary, buying the stocks of low past growth rate leads to higher subsequent returns
131
since over a period of time investors become aware that they have been too pessimistic in
predicting future growth and the stock price rises.
Another reason for the existence of value premium is the mean reversion tendency
of stocks. According to Graham and Dodd (1934) the firms who have and are currently
experiencing high (low) earnings growth are unlikely to be able to sustain it to the extent
expected by the market. When this earnings growth reverts towards some
industry/economy-wide mean, then this will result in a revision of the earnings
expectations, a fall in the firm's price-to-earnings multiple and so a downward correction
in its stock price (Bird and Gerlach, 2003). Also some authors believed that the value
premium exists due to random ocuurences i.e. the instances that are not likely to happen
again in the near future (Lo and MacKinlay, 1990; Breen and Korajczyk, 1995; Bauamn
et. al., 1998). Thus, the behavioral school of thought is conjectured to be the reason
behind the existence of value premium in Indian stock market.
4.5 DETERMINING THE NUMBER OF PRINCIPLES A PARTICULAR

STOCK SHOULD FOLLOW TO RENDER IT A GOOD INVESTMENT
The important question is if the investors have to decide whether to buy a
particular stock or not on the basis of Graham‟s criteria, then they have to examine each
stock against each principle. In that situation they need to assess how many principles are
fulfilled by a particular stock and how many principles in totality a particular stock
should follow so as to render it as a good value stock. Therefore, an effort has been made
to examine each and every principle individually as well as collectively. All the
principles have been evaluated by forming individual (binary digits) score as well as
composite score for all the companies in the sample. Every company has been examined
for all the ten principles or rules of Benjamin Graham. If the stock meets the particular
rule, it is given score 1 and otherwise 0 and then the scores of all the principles for a
particular stock are totaled to calculate the composite score. If a company meets 3 rules,
then it will be given a score of 3 out of 10 and if it meets all the rules, it will be given a
score of 10. Hence, the aggregate score is the sum of individual binary signals (Piotroski,
2000). Thus, total score acts as scale variable which helps one to evaluate the number of
rules/ principles each company has met.
132
Graham‟s principles
Rules P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 Total
Co. score
A 1 0 1 1 0 1 0 0 0 1 5
B 0 1 0 0 1 0 1 0 0 0 3
C 1 1 1 1 1 1 0 1 1 1 9
D 0 0 0 0 0 0 0 0 0 0 0
E 1 1 1 1 1 1 1 1 1 1 10
Where 1= fulfillment of the rule; 0= non fulfillment of the rule
Figure 4.1: Computation of total score
Each and every stock has been examined for all the principles to see whether they
fulfill the particular principle or not and how many principles in total a particular stock
follows. Table 4.19 shows the score wise distribution of the sample companies across the
period of study.
Table 4.19 shows that across the period of 15 years, not even a single company has been
found that has met all the principles of Graham‟s stock selection criteria. Also, only two
companies have been able to fulfill nine rules of the criteria. The maximum number of stocks
is clustered around the total score of 4. Therefore, it clearly confirms the stringency of these
principles.
4.5.1 Analyzing the Market Adjusted Performance of Different Categories of Total
Score
Further, the market adjusted returns of all the stocks across the period of the study having
the total score of 0, 1, 2,…, 10 is calculated and analyzed using one sample t-test. The
following hypothesis is examined:
H015 : The mean/median market adjusted return of stocks clubbed into
different categories of total score is equal to zero.
133
Table 4.19: Distribution of Stocks Amongst Different Categories of Total Score Across the Period of Study (1996- 2010)
Total score Total no. of

1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 stocks
0 1 0 1 0 0 0 0 1 0 1 1 0 0 1 2 8
1 0 4 0 2 4 2 3 2 7 9 13 9 9 8 24 96
2 17 11 3 8 9 10 20 5 10 24 34 49 52 23 61 336
3 13 15 10 19 10 14 15 23 14 42 38 54 59 47 54 427
4 21 13 18 15 28 23 19 22 39 42 39 43 51 46 51 470
5 9 10 15 12 10 15 16 20 22 13 19 17 22 38 38 276
6 4 7 14 10 13 16 9 12 17 9 6 7 14 25 15 178
7 1 3 5 5 6 9 5 9 10 3 3 3 4 14 5 85
8 0 2 2 4 6 4 0 2 0 0 1 1 1 6 2 31
9 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 2
10 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Total no. of
stocks 66 65 68 75 86 94 87 96 119 143 154 183 213 208 252 1909
134
The results are reported in the Table 4.20 below:

Table 4.20: Significance of Market Adjusted Returns of the Stocks Classified into
Different Categories of Total Score
Total
No. Mean Mean
Score Std. Dev. t-value p-value Std. Dev. t-value p-value
(Annual) (Annualized)
11.0207 -10.5775
0 8 42.70422 - .389 40.78363 - .547
(15.09822) (14.41919)
-12.1699 -8.6519
1 96 47.64483 -2.503 .014** 31.29960 -2.708 .008***
(4.86273) (3.19450)
-1.2759 -2.5652
2 336 43.98545 -.532 .595 30.95905 -1.519 .130
(2.39960) (1.68895)
1.2292 .7672
3 427 46.92201 .541 .589 29.00795 .547 .585
(2.27072) (1.40379)
6.5662 2.5900
4 470 50.03010 2.845 .005*** 29.00748 1.936 .054*
(2.30772) (1.33802)
8.7960 4.6742
5 276 48.07244 3.040 .003*** 28.71880 2.704 .007***
(2.89362) (1.72867)
27.5540 13.0276
6 178 54.17062 6.786 .000*** 31.83039 5.460 .000***
(4.06026) (2.38579)
26.4221 12.1192
7 85 54.93403 4.434 .000*** 29.08219 3.842 .000***
(5.95843) (3.15441)
40.3791 22.3236
8 31 80.96885 2.777 .009*** 33.86981 3.670 .001***
(14.54243) (6.08320)
37.3456 26.5782
9 2 61.11984 - .467 22.12193 - .392
(43.21825) (15.64257)
10 0 - - - - - - - -
Note:
2. Standard error of the mean has been reported in parenthesis
It is evident from Table 4.20 that the number of stocks that has a total score of 0 is 8.
The number of stocks which get total score of 1 is 96. Similarly, the number of stocks that are
able to fulfill 9 principles of Graham is only 2. Not even a single stock has been found across
the period of 15 years which is able to fulfill all the rules. Thus, the stringency of the Graham
principles to be followed by stocks in present scenario goes without saying. It is also
interesting to note that from total score 0 till total score of 3, the market adjusted returns of
the stocks in case of 12 months holding period have been either insignificant or negative.
From the total score of 2 till the total score of 9, the mean market adjusted returns increase
when there is one point improvement in the total score. Thus, higher the total score, larger the
returns. From the total score of 4 till total score of 8, the mean returns have been significantly
higher than the market.
135
Similarly, when the holding period of the stocks has been extended from 12 months
to 24 months, we find that the market adjusted return has not been significant at 5% level of
significance till the stocks have the total score of 4. Thus, when the stocks have the total score
5 or more, their mean market adjusted returns become significantly positive at 1% level of
significance. Thus, the stocks fulfilling at least five rules of Graham are considered to be a
good investment.
4.5.2 Analyzing the Difference in Mean Returns of High Scoring and Low Scoring
Stocks
Further analysis corresponds to analyzing the significance of difference between the
mean market adjusted return of stocks having high score and the stocks having low score.
The independent sample t-test has been employed on the total sample of stocks classified into
2 categories on the basis of total score i.e. stocks getting high total score and the stocks
getting low total score. The difference of returns between the two, if proved significant, will
show that the stocks meeting maximum rules of Graham will undoubtedly outperform the
stocks meeting minimum or no rule. In order to apply independent sample t-test, the firms
are then divided into two groups i.e. the firms which have lowest aggregate score (score
equals 0 or 1) and the firms which have highest aggregate score (score equals 8 or 9). Along
with it, the t-test has also been used to examine the return differences among the firms with
aggregate score (0, 1 or 2) and (7, 8 or 9), the firms with aggregate score (0, 1, 2 or 3) and (6,
7, 8 or 9) and the firms with aggregate score (0, 1, 2, 3 or4 ) and (5, 6, 7, 8 or 9). It will thus
help to know whether high score firms differ significantly in returns from the low score
firms. The following hypothesis is examined:
H016 : There is no significant difference in the mean market adjusted returns of

the stocks that fulfill maximum rules of the criteria and the stocks that
fulfill minimum rules.
Table 4.21 shows the results of t-test.
136
Table 4.21: Results of t-test Employed on Market Adjusted Returns of the Firms Having High Score Compared to the
Firms Having Low Score
F-value
Assumption Assumption
F-value of of
Difference of Difference of
No. of Levene’s Levene’s
S. No. Categories Mean in mean Equal in mean Equal
stocks Std. test of Mean returns Std. test of
returns return of Variances t-value return of Variances t-value
Dev. equality of (Annualized) Dev. equality
(Annual) two /No Equal two /No Equal
variance of
groups Variances groups Variances
variance
Assumed Assumed
High
572 19.06 54.08 9.41 30.37
(5,6,7,8,9) No equal No equal
1 17.49 4.684** 6.705*** 9.58 0.055 6.39***
Low variance variance
1337 157 47.55 -0.173 29.87
(0,1,2,3,4)
High
296 28.72 57.67 13.78 31.29
(6,7,8,9) No equal No equal
2 29.85 10.982*** 8.06*** 15.45 0.061 7.51***
867 -1.13 45.96 -1.67 30.24
(0,1,2,3)
High
118 29.73 62.58 14.87 30.53
(7,8,9) No equal No equal
3 33.16 11.185*** 5.37*** 18.91 0.015 5.84***
440 -3.42 44.95 -4.03 31.25
(0,1,2)
High
33 41.91 79.78 23.56 33.07
(8,9) equal No equal
4 52.29 7.177 3.52*** 32.36 0.093 4.977***
Low 104 -10.38 47.49 variance -8.8 31.88 variance
(0,1)
137
It can be seen from Table 4.21 that in case of 12 months holding period of stocks,
the mean market adjusted return difference of 52.29% between the firms with high score
(8, 9) compared to firms with low score (0, 1) is statistically significant at 1% level of
significance. Similarly, when the mean market adjusted return difference between the
firms scoring 7, 8 or 9 and the firms scoring 0, 1 or 2 is examined, the same turned out to
be significant at 1% level of significance. Similar results are found when the firms with
the score of 6, 7, 8 or 9 are compared with the firms with the score of 0, 1, 2 or 3 and the
firms with the score of 5, 6, 7, 8 or 9 with the firms having total score 0, 1, 2, 3 or 4 on
the basis of market adjusted return. It therefore leads to the rejection of null hypothesis
(H016) of no significant difference in the mean market adjusted returns of the stocks that
fulfill maximum rules of the criteria compared to the stocks that fulfill mimimum rules.
Thus, high score firms clearly outperform low score firms on the basis of stock return
entailing a fairly monotonic positive relationship between high score and subsequent
returns. However, the F-value of the Levene‟s Test intended to examine the variances of
two groups (low score group and high score group), is significant at 5% level of
significance in all the cases [except high score (8, 9) compared to low score (0, 1)]
showing that the variances of two groups are not equal. Thus, the results of Welch t-test
become applicable in case of 12 months holding period.
Table 4.21 also reports the results of independent sample t-test employed to
examine differences in the groups of firms with high score and firms with low score in
case of 24 months holding period of stocks. It is apparent from Table 4.21 that the
difference of 9.58% in the mean market adjusted returns of the firms with high score (5,
6, 7, 8 or 9) compared to firms with low score (0, 1, 2, 3, 4) is significant at 1% level of
significance. It shows that the firms that fulfill maximum rules of Graham‟s stock
selection methodology outperform the firms that fulfill no or minimal rules. Similar
results have been found when the returns of firms with high score (6, 7, 8 or 9) are
compared with the returns of firms with low score (0, 1, 2 or3). Likewise, in other 2 cases
also, high score firms outperform low score firms. The F-value of the Levene‟s Test
intended to examine the variances of two groups (low score group and high score group),
is insignificant in all the cases showing that the variances of two groups are equal. Thus,
138
extending the holding period tend to equalize the variances in two groups (Malkiel,
2012).
4.5.3 Analyzing the Presence of Abnormal Returns Yielded By Different

Categories of Total Score
We proceed to examine the abnormal returns in case of stocks classified into
different categories of total score, for which the following hypothesis is tested:
H017 : The abnormal return yielded by the stocks clubbed into different
categories of total score is equal to zero.
The results of asset pricing model used to assess the presence of abnormal returns
amongst different categories of total score in 12 months holding period are reported in
Table 4.22
Table 4.22: Results of Asset Pricing Model Applied on Stocks Grouped into Different
Categories of Total Score in case of 12 Months Holding Period
2
Total R DW ANOVA Results Alpha Beta
score F- P-value Coefficients t- p-value Coefficients t- p-value
value value value
1 - - - - - - - - - -
2 0.0168 1.94 2.931 0.088* .821 1.057 .291 .168 1.712 .088*
(.776) (.098)
3 0.044 1.929 9.410 .002*** .401 .555 .578 .280 3.067 .002***
(.722) (.091)
4 0.027 1.851 4.981 0.027** 1.017 1.523 .130 .188 2.232 .027**
(.668) (.084)
5 0.036 1.944 6.553 0.011** 1.148 1.496 .137 .249 2.560 .011**
(.768) (.097)
6 0.030 1.782 5.434 0.021** 2.277 2.662 .008*** .252 2.331 .021**
(.855) (.108)
7 0.053 2.176 10.098 .0017*** 2.070 2.275 .024** .365 3.177 .001***
(.909) (.115)
8 - - - - - - - - - -
9 - - - - - - - - - -
Note:
2. Standard error of the coefficients has been reported in parenthesis
3. The stocks pertaining to total score 1, 8 and 9 have been zero in certain years, due to which
the time series data for those years goes missing. Hence, the time series CAPM model could
not be applied on those scores.
Table 4.22 provides the results of asset pricing model applied on the stocks falling
into the different categories of total score. The F-value of the ANOVA is significant in all
139
the categories of total score showing the fitness of the regression model. The value of
beta has been lesser than 1 in all the cases showing the lesser volatility of the portfolios
than the market portfolio. Table 4.22 also shows that the stocks having total score of 6, 7
have significant alpha showing that if a stock that is able to fulfill at least any 6 rules of
Graham, is able to provide abnormal returns to the investors in case of 12 months holding
period. Thus, if the particular stock fulfills lesser than 6 rules of Graham, such a stock
cannot yield statistically significant excess returns to investors in case of 12 months
holding period. The results of 24 months holding period are as follows:
Table 4.23: Results of Asset Pricing Model Applied on Stocks Grouped into Different
Categories of Total Score in case of 24 Months Holding Period
Total
R2 DW F- Coefficients t- p-value Coefficients t- p-value
score p-value
value value value
1 - - - - - - - - - -
.710 .207
2 0.025 1.953 9.03 0.003*** 1.313 .190 3.005 .003***
(.541) (.069)
.814 .180
3 0.021 1.945 7.596 0.006*** 1.587 .113 2.756 .006***
(.513) (.065)
1.000 .192
4 0.028 1.939 10.171 0.002*** 2.120 .035** 3.189 .002***
(.472) (.060)
1.306 .253
5 0.034 1.918 12.733 0.000*** 2.352 .019** 3.568 .000***
(.555) (.071)
1.825 .262
6 0.032 1.877 11.668 0.001*** 3.033 .003*** 3.416 .001***
(.602) (.077)
1.319 .264
7 0.025 2.059 9.071 0.003*** 1.921 .056* 3.012 .003***
(.687) (.088)
8 - - - - - - - - - -
9 - - - - - - - - - -
Note:
2. Standard error of the coefficients has been reported in parenthesis.
3. The stocks pertaining to total score 1, 8 and 9 have been zero in certain years, due to which
the time series data for those years goes missing. Hence, the time series CAPM model could
not be applied on those scores.
Table 4.23 shows the results of asset pricing model when the holding period of
the stocks clubbed into different categories of total score is extended from 12 months to
24 months. The F-value of the ANOVA is significant in all the cases showing that the
model is fit. The value of beta has been lesser than 1 in all the cases showing the lesser
volatility of the portfolios than the market portfolio. The value of alpha is significant in
140
case of stocks having total score of 4, 5, 6 and 7. It shows that the stocks that fulfill any 4
rules (at least) of Graham‟s stock selection criteria can provide significantly excess
returns to investors if they hold such portfolio for the period of 24 months.
The explanatory power of CAPM model in all the cases has been very low. Thus,
the market risk factor has little role in explaining the variation in overall returns. Along
with market risk factor, there are other risk factors also, that affect the stock returns. Two
most popular risk metric devised in number of studies (Fama and French, 1996; Sehgal,
2001; Malin and Veeraraghavan, 2004; Bahl, 2006; Bundoo, 2008) that affect stock
returns are size factor (stocks with smaller market capitalization outperform the stocks
with larger market capitalization) and the value factor (stocks with higher book to market
ratio outperform the stocks with lower book to market ratio). In the present study, the
number of stocks meeting the specific criterion are very small in certain cases (principle
of price being lesser than two- third the net current asset value, refer Table 4.5). Thus, the
calculation of risk factors i.e. size, value is not possible in that situation. Therefore, we
have focused upon the opportunity cost of capital as well as market risk factor in
examining the abnormal performance of the principles. However, in order to check the
robustness of the Graham‟s stock selection rules, we have incorporated size and value
framework in panel data format. Thus panel data regression will help to assess the
predictive capability of the total score in explaining the stock return.
4.6 ANALYZING THE PREDICTIVE ABILITY OF THE STOCK

SELECTION CRITERIA
The above results make it evident that the total score can be used as a tool to
discriminate between the firms with strong future return performance and the firms with
weak subsequent performance (refer Table 4.21). This section aims to explore the role of
total score in explaining risk adjusted returns. A brief explanation of the all the variables
incorporated in the study are given below.
4.6.1 Model Development

4.6.1(A) Dependent variable
In order to examine the predictive ability of the criteria in explaining the stock
returns, the market adjusted stock return (annual in case of 12 months, annualized in case
141
of 24 months) variable has been taken as dependent variable as advocated in Piotroski

(2000), Michou (2007), Dahl et al. (2009) and Dosamantes (2013). The reason behind
taking excess of portfolio‟s return over market is to know whether the excess returns
yielded by the stocks are explained by total score or not.
4.6.1 (B) Explanatory Variables and Hypothesis Development

A brief discussion of the explanatory variables used in the model is given below.
Total score - The model estimates the relation between the total score and market
adjusted returns. The total score has been calculated on the basis of number of principles
met by each stock as described in Section 4.5. Therefore, to examine the role of total
score in explaining market adjusted returns, the following hypothesis is tested.
H018 : Total score of the stocks has no significant impact on the market
adjusted returns.
Size - The term „size‟ is measured as the market value of the equity shares (a stock‟s
price times shares outstanding). Banz (1981) found that the size factor significantly add
to the explanation of cross section of average stock returns. Fama and French (1992) also
found significantly negative impact of size on stock returns, thereby implying that the
stocks with smaller size have higher risk adjusted returns than the stocks with larger size.
In addition, Mukherji et al. (1997), Anderson et al. (2003), Dunis and Reilly (2004),
Kumar and Sehgal (2004), Kyriazis and Diacogiannis (2007), Tripathi (2009) have also
found significantly negative impact of size on stock returns. Thus, the variable „size‟ acts
as control variable in order to estimate the relationship between market adjusted returns
and the total score. The following hypothesis is formulated and tested.
H019 : The size of the stock has no significant impact on its market adjusted
returns.
Book to market effect- It is measured as the ratio of the book value of the company to
number of shares outstanding. Lakonishok et al. (1991) found that the book to market
(B/M) ratio has the strong role in explaining the cross section of average stock returns.
Fama and French (1992) also found the significant impact of book to market effect on
stock returns. In addition, Sehgal (2001), Malin and Veeraraghavand (2004), Bahl (2006)
142
and Bundoo (2008) have also found that the book to market effect significantly add to the
explanation of cross section of average stock returns. Thus, book to market ratio acts as
control variable in order to estimate the relationship between market adjusted returns and
the total score. The following hypothesis is formulated and tested.
H020 : The book to market equity of the stock has no significant impact on its
market adjusted returns.
4.6.2 Model Estimation

As the present data set entails both a spatial (cross sectional units i.e. companies)
and temporal dimension (periodic observations of a set of variables characterizing these
cross-sectional units over a particular time span); thus to examine these issues, panel data
analysis has been used. The following model is examined:
Marketadjustedreturns 1
totalscoreit 2
log sizeit 3
booktomarketit uit
2
Here, uit ~ N 0, u , i 1,..., N (N= no. of cross-sectional units), and t 1,..., T
(T= no. of time-series units).

Before conducting regression analysis, the association between dependent
variable and explanatory variables and also amongst different explanatory variables has
been examined using correlation analysis. Table 4.24 provides the correlation matrix for
market adjusted returns (dependent variable) and 3 selected explanatory variables for the
period 1996-2010.
Table 4.24: Pearson’s Product Moment Correlation Matrix
24 months
12 months market
market adjusted
Size B/M ratio Total score adjusted returns
returns
(annual)
(annualized)
Size 1 -.322*** -.284*** -.116*** -.114***
B /M ratio -.322*** 1 .191*** .136*** .138***
Total score -.284*** .191*** 1 .197*** .186***
12 months market -.116*** .136*** .197*** 1
adjusted returns
(annual)
24 months market -.114*** .138*** .186*** 1
adjusted returns
(annulized)
143
Table 4.24 shows that the dependent variable (market adjusted returns) has high
degree of correlation with the explanatory variables in both the holding periods. Also,
significantly negative correlation has been observed between the market adjusted returns
and the size factor (-0.116 in case of 12 months; -0.114 in case of 24 months holding
period). It implies that increase in size results in fall in market adjusted returns. In
addition, total score has significantly positive association with the market adjusted
returns (0.197 in case of 12 months; 0.186 in case of 24 months holding period) which
implies that as the value of total score increases, there would be an increase in market
adjusted returns As far as correlation amongst the explanatory variables (size, book to
market ratio, total score) is concerned, it is lesser than the prescribed rule of thumb i.e.
0.8 (Gujarati, 2007). We further proceed to apply regression analysis.
The results of panel data regression to assess the predictive capability of the total
score in explaining the stock return are shown in Table 4.25 as under:
Table 4.25: Results of Pooled OLS Regression where Market Adjusted Return has
been taken as Dependent Variable and the Total Score as Independent
Variable along with size and B/M ratio as Control Variables.
Dependent Market adjusted returns in case Market adjusted returns in case of
variable of 12 months holding period 24 months holding period
Coefficients Robust z- Coefficients Robust z -value
Std. Err. value Std. Err.
Independent Total score 5.251559 .7136504 7.36*** 2.924985 .462802 6.32***
variables
Log (Size) -1.123328 .8586486 -1.31 -.6491631 .420047 -1.55
B/M ratio .8720236 .5936261 1.47 .5404193 .2252618 2.40**
constant -3.6432 10.1589 -0.36 -2.945508 5.24015 -0.56
Chow test- 1.21 0.85
F-stats (0.2621) (0.6104)
(P-value)
LM Test
Chi-square 1.21 0.00
stats (0.2621) (1.000)
(P-value)
R2 0.0502 0.04667
Table 4.25 shows that the value of chow test, LM test has been insignificant in
both the holding periods showing the absence of fixed, random effects in the data. Hence
pooled OLS (ordinary least square) regression has been used. Further, the coefficient of
144
size factor (log of market capitalization of shares) is negative in both the periods showing
that that larger the market capitalization of shares, lesser would be their market returns.
Thus, stocks with smaller market capitalization outperform the stocks with larger market
capitalization. However, the insignificant Z-value of the size variable leads to the
acceptance of null hypothesis (H019) of no significant impact of size on market adjusted
returns. Also, the coefficient of book to market ratio is positive stating that higher the
book to market ratio of the stocks, greater will be the market returns. Also, the book to
market variable has been significant at 5% level of significance in explaining the
variation in overall returns in case of 24 months holding period. It leads to the rejection
of null hypothesis (H020) of no significant impact of book to market ratio on market
adjusted returns.
In respect of total score we notice that, after controlling for size and book to
market effect, the coefficient of total score is positive and significant at 1% level of
significance in both the holding periods leading the rejection of null hypothesis (H018) of
no significant impact of total score on market adjusted returns. It therefore implies that 1
point improvement in total score will result in 5.25 % increase in market adjusted stock
returns in case of 12 months holding period and about 2.93% increment in market
adjusted returns in case of 24 months holding period, thus clearly depicting the
importance of total score in explaining the stock returns.
Since the total score is the sum of the ten individual scores; so there is a
possibility of some principles affecting stock returns positively and some negatively.
Also, some principles might not be affecting stock returns at all. Therefore, a need was
felt to examine the predictive ability of each and every principle in explaining the stock
returns. The following model is examined:
Marketadjustedreturns 1
P1it 2
P 2it 3
P3it 4
P 4it 5
P5it 6
P 6it
7
P 7it 9
P9it 10
P10it 11
log sizeit 12
booktomarketit uit
Before conducting regression analysis, the association between dependent

variable and explanatory variables and also amongst different explanatory variables has
been examined using correlation analysis. Table 4.26 provides the correlation matrix for
market adjusted returns (dependent variable) and selected explanatory variables for the
period 1996-2010.
145
Table 4.26: Pearson’s Product Moment Correlation Matrix

12
24 months
months
market
B/M market
Size P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 adjusted
ratio adjusted
returns
returns
(annualized)
(annual)
Size 1 -.322*** -.388*** -.246*** -.221*** -.600*** -.274*** .189*** -.056** .218*** .079*** .166*** -.116*** -.114***
B/M ratio -.322*** 1 .208*** .126*** .179*** .321*** .209*** -.097*** -.016 -.078*** -.034 -.032 .136*** .138***
P1 -.388*** .208*** 1 .304*** .330*** .477*** .160*** -.096*** .046** -.165*** .068*** .004 .195*** .150***
P2 -.246*** .126*** .304*** 1 .111*** .258*** -.007 -.026 .021 -.113*** .115*** .028 .079*** .061***
P3 -.221*** .179*** .330*** .111*** 1 .329*** -.003 -.003 -.017 -.108*** -.071*** -.052** .099*** .070***
P4 -.600*** .321*** .477*** .258*** .329*** 1 .321*** -.138*** -.024 -.203*** -.130*** -.158*** .144*** .155***
P5 -.274*** .209*** .160*** -.007 -.003 .321*** 1 .063*** -.068*** .229*** -.015 -.028 .135*** .133***
P6 .189*** -.097*** -.096*** -.026 -.003 -.138*** .063*** 1 -.176*** .304*** .070*** .073*** .071*** .057**
P7 -.056** -.016 .046** .021 -.017 -.024 -.068*** -.176*** 1 -.104*** -.042* -.031 -.002 .005
P8 .218*** -.078*** -.165*** -.113*** -.108*** -.203*** .229*** .304*** -.104*** 1 .077*** .114*** .050** .056**
P9 .079*** -.034 .068*** .115*** -.071*** -.130*** -.015 .070*** -.042* .077*** 1 .427*** .015 .020
P10 .166*** -.032 .004 .028 -.052** -.158*** -.028 .073*** -.031 .114*** .427*** 1 .038* .038*
12 months
market
adjusted -.116*** .136*** .195*** .079*** .099*** .144*** .135*** .031 -.002 .050** .015 .038* 1
returns
(annual)
24 months
market
adjusted -.114*** .138*** .150*** .061*** .070*** .155*** .133*** .057** .005 .056** .020 .038* 1
returns
(annualized)
146
Table 4.26 makes it evident that the explanatory variables; book to market ratio,
P1, P2, P3, P4, P5, P6, P8 and P10 have significantly positive correlation with the
dependent variable (market adjusted returns) in both the holding periods. Also,
significantly negative correlation has been observed between the market adjusted returns
and the size factor (-0.116 in case of 12 months; -0.114 in case of 24 months holding
period). It implies that increase in size results in fall in market adjusted returns. Also, P7
has been negatively associated with the returns which implies that the fulfillment of
principle of current ratio being greater than two, leads to lesser market adjusted returns
in case of 12 months period. In case of 24 months period, the association between the
returns and P7 has been insignificant. In addition, very low degree of correlation has
been found between the returns and the principle of growth in earnings (P9) in both the
holding periods. As far as correlation amongst the explanatory variables is concerned, it
is lesser than the prescribed rule of thumb i.e. 0.8 (Gujarati and Sangeetha, 2007). We
further proceed to apply regression analysis.
Table 4.27 shows the panel data regression results wherein market adjusted
return act as dependent variable and the ten principles act as ten independent variables6
along with size and book to market ratio as control variables.
Table 4.27: Results of Pooled OLS Regression wherein Market Adjusted Return Act
as Dependent Variable and the Ten Principles Act as Ten Independent
Variables along with Size and book to Market Ratio as Control Variables
Dependent Market adjusted returns in case Market adjusted returns in case of
variable of 12 months holding period 24 months holding period
Coefficient Robust z-value Coefficient Robust z -value
Std. Err. Std. Err.
Independent P1 15.97838 3.107795 5.14*** 5.894766 1.789278 3.29***
variables
P2 2.135174 2.744743 0.708 .2047919 1.628215 0.13
P3 4.213673 3.570884 1.18 .1200493 1.921258 0.06
P4 2.80843 3.660787 0.77 5.455953 2.008602 2.72***
P5 15.9304 7.965386 2.00** 7.437202 4.190464 1.77*
P6 6.07784 4.579964 1.33 6.972771 2.985041 2.34**
P7 1.074026 2.321659 0.46 1.672771 1.286448 1.31
P8 6.556941 2.303775 2.85*** 3.973161 1.484429 2.68***
P9 -1.317627 2.330512 -0.57 .1285186 1.467212 0.09
6
The ten principles act as dummy variables whereby the stock that fulfills a particular principle is
given a score of 1 and a stock which does not fulfill a particular principle is given a score of 0.
147
P10 4.790476 2.464487 1.94* 2.889588 1.645647 1.76*

Log (Size) -.4091686 .7961032 -0.51 -.294346 .4281022 -0.69
BM ratio .7621655 .511878 1.49 .5028986 .1710606 2.94***
constant -8.826557 9.987545 -0.88 -9.43206 5.64616 -1.67*
Chow test 1.12 0.82
F-stats (0.3372) (0.6475)
(P-value
LM Test
Chi- 0.00 0.00
square (1.000) (1.000)
stat
(P-value)
R2 0.0648 0.0576
Table 4.27 shows that the value of chow test, LM test has been insignificant in both the
holding periods showing the absence of fixed, random effects in the data. Hence pooled
OLS (ordinary least square) regression has been used.
From Table 4.27 we further notice that the principle of earnings yield (P1) has
been a significant factor in both of the holding periods in explaining variation in overall
returns. Therefore, larger the earnings yield a firm has, larger the returns it will deliver.
Further, the principle of price shrinkage (P2) has been an insignificant variable in
explaining the overall stock returns. Also, the principle of high dividend yield (P3) has
not been a significant variable. This could be due to the fact that due to larger emphasis
on growth in recent years, the perception regarding the valuation of one unit of retained
earnings lesser than a unit of dividends has been changed in the minds of investors and
management. Several past surveys of shareholder opinion indicate that earnings and
capital gains do, in fact, weigh more heavily than dividends in evaluating the relative
desirability of alternative stock investments (Friend and Puckett, 1964). Moreover, while
searching for investment opportunities, a firm seeking growth is also likely to invest in
substantial proportion in research and development expenses. Consequently, for a firm,
successful investment opportunities are likely to result, over time, in the buildup of the
firm's assets. Thus, the firm‟s ability to pay dividends is constrained by the requirement
of investment outlays for positive net present value (NPV) investment projects (Reddy
and Rath, 2005). Thus, given the firm‟s investment and financing decisions, a small
dividend payment corresponds to high earnings retention with less need for externally
generated equity. Also, after 1980s, the dividend yields declined as management
148
attempted to free up cash, for reinvestment to sustain the long-term earnings power of the
company (Tengler, 2003).
Further, the principle of low price in relation to the tangible book value (P4) has
been a significant factor in explaining the overall returns in case of 24 months holding
period. Along with it, the principle of low price in relation to the net current assets of the
company (P5) is also a significant factor in explaining the overall returns. Thus, the
behavioral school of thought (as detailed in section 4.4) could be considered responsible
for yielding high returns through buying these undervalued stocks.
The principle of debt being lesser than book value (P6) and the principle of debt
being lesser than twice the net current asset value (P8) is significant in both the holding
periods showing that the firms that rely lesser on the borrowed funds, have higher risk
adjusted returns. Therefore, the importance of leverage and the positive impact of
leverage on stocks returns as documented in various studies like, Hamada (1972), Ross
(1977), Heinkel (1982), Marsh (1982), Bhandari (1988), Mukherji et al. (1997), Ooi
(1999), Fama and French (2002), Lasher (2003), Ding et al. (2005), Dhaliwal et al.
(2006), Ward and Price (2006), Sharma (2006) and Tripathi (2009), could not be held in
this study.
However, different researchers have totally negated the larger reliance on debt
saying that the managers of the firms with valuable growth opportunities should choose
lower leverage because the firms might not be able to take advantage of their investment
opportunities if they have to raise outside funds. Increased leverage reduces both current
funds available for investment as well as the firm‟s ability to raise additional funds to
invest. Also, the firms with poor growth opportunities should be prevented from
dissipating cash flows on poor projects. The different studies, such as, Myers (1977),
Stultz (1990), Harris and Rajiv (1991), have found that in extreme cases, a firm‟s debt
overhang can be large enough to prevent it from raising funds to finance positive net
present value projects. Also, Kortewerg (2004) stated that net benefit of leverage is high
for highly profitable firms with low depreciation, stable profits, and low P/B ratios and in
the period of economic expansions. Severely distressed firms have high distress cost in
case of high debt. The distress cost represents the present value of the future cash flows
that are lost due to the presence of debt in the firm‟s capital structure. Further, Titman
149
and Wessels (1988), Dimitrov and Jain (2005), Penman et al. (2007), Muradoglu and
Sivaprasad (2008), George and Hwang (2009) also reported the negative relation between
the two. Myers (1984) stated that successful companies need not to place much
dependence on external funding, since they rely on internal reserves. Further, Myers
(2001) stated that high level of debt increases the possibility of bankruptcy. Fama and
French (1988) also asserted that more profitable firms tend to have lower level of debt.
Increasing debt would signal poor future prospects of firms since future earnings will be
impacted negatively due to cash flow being used to service debt reducing the amount of
money available for fund future development.
Thus, greater the firm's leverage, higher the variability of the earnings available
for distribution to shareholders. In other words, as the firm's debt-equity ratio increases,
the risk borne by the shareholder increases and the expected utility of the investment
decreases. Thus, the required return from investment must increase in order to retain its
appeal to the shareholders (Arditti, 1967).
Also evident from Table 4.27 that the principle of current ratio being greater than
two (P7) is insignificant factor in explaining the variation in market adjusted returns in
Indian stock market. Thus, higher current ratio is not desirable in present market scenario.
Different studies have observed the negative relationship between the working capital
and the profitability opining that shortening cash conversion cycle always improves
profitability i.e. the excess inventory leads to low profitability [see, for example, Soenen
(1993), Shin and Soenen (1998), Wang (2002), Deloof (2003), Eljelly (2004), Lazaridis
and Tryfonidis (2006), Padachi (2006), Raheman and Nasr (2007), Garcia-Teruel and
Martinez-Solano (2007), Ganesan (2007), Azhar and Saad (2010), Gill et. al (2010),
Dong and Su (2010), Danuletiu (2010), Izadinia and Taki (2010), Hayajneh and Yassine
(2011), Vahid et. al (2012)]. It may be due to the reason that a high current ratio implies
heavy investment in current assets which may mean under utilization of resources or
more liquidity than required which leads to lesser profitability (Chary, et. al, 2011). A
high current ratio which used to be a safety measure earlier reveals only blocked funds in
briskly growing economy and therefore impacts stock returns negatively. Moreover, this
ratio does not distinguish between different types of current assets, some of which are far
150
more liquid than others. A company could be getting into cash problems and still have a
strong current ratio (Walsh, 1993).
The earnings growth variable (P9) has been insignificant in both the holding
periods suggesting that the high growth in earnings variable is not able to explain the
variation in the stock price. Thus, the naive strategy of investing in the high growth
shares is not said to be working in the Indian stock market. Therefore, the forecasters
who rely too heavily on past trends while forming their expectations about the future
leads to biased estimates of future equity returns (Bauman and Miller, 1997). Since the
fundamental variables have mean reversion tendency, firms who have and are currently
experiencing high (low) earnings growth are unlikely to sustain it to the extent expected
by the market. When this earnings growth reverts towards industry/ economy mean, then
this will result in the revision of earnings‟ expectations (Bird and Gerlach, 2003). Thus,
the firms with low growth rate in earnings tend to outperform the firms with high growth
rate as evidenced in Bauman and Miller (1997), Bauman et al. (1998), Ahmed and Nanda
(2001).
Stability in earnings (P10) is the significant variable explaining the variation in
returns. Stocks with more stable earnings tend to be underpriced by markets due to the
fact that company with stable earnings are often boring companies that do not make the
news and the passionate investors are not interested in them. As a result, the companies
with stable earnings will be underpriced relative to companies with more volatile
histories (Damodaran, 2004). Also, the earnings surprise i.e. the difference between
earnings estimates and the actual announced earnings affects a stock‟s volatility and
returns. An earnings surprise is significantly correlated with volatility and overnight
returns (Lim, 2009). There is high volatility in stock prices around the earnings
announcement. Prices behave significantly different around this particular event
conforming that these changes are due to announcement of earnings (Qureshi et al.,
2012). However, greater volatility associated with the event renders them more risky.
Therefore, in such a situation the companies having smoother earnings have lesser
earnings surprise for investors and thus are lesser risky than the companies with more
volatile earnings. Investors investing in such companies are neither too pessimistic nor
151
too optimistic about the future scenario. Thus, on account of being ignored in the market,
such companies yield higher risk adjusted returns to investors in long term.
4.7 CONCLUSION
The study examines the profitability of Benjamin Graham‟s criteria of stock
selection based on value investing principles in Indian stock market. The rules have been
formed on the core principle of value investing i.e. buying a stock when its intrinsic value
is greater than its market price. Each and every stock selection rule of Graham holds the
essence of basic phenomenon of value investing. The ten principles when examined
individually did not outperform the market every year. However, across the period of
study i.e. 15 years, each and every principle has yielded higher average return than that of
market return. On the other hand, when the statistical significance of the mean returns of
the stocks that fulfill the principles and the stock which do not fulfill the principles is
examined, the same turned out to be insignificant in case of stocks having higher book
value in relation to debt, current ratio greater than two and the stocks which have shown
the growth in their per share earnings over a period of time in case of one year holding
period of such stocks. Therefore, one becomes skeptical regarding the profitability of a
few principles of Graham‟s stock selection criteria. However, when the holding period of
the stocks is extended to two years, the mean returns of the stocks that have higher book
value in relation to debt are found to be significantly higher than the stocks that have
higher debt in relation to book value. Further, the abnormal returns yielded by these
principles are estimated using asset pricing model, which showed that the abnormal
returns in case of stocks having high dividend yield, high book value in relation to debt
and current ratio greater than two, are not present in case of one year holding period.
When the holding period of the stocks was extended to two years, the model reconfirmed
the absence of abnormal returns in case of stocks having high dividend yield. Thus, in
order to finance the future growth opportunities, the firms goes for lesser dividend
payments and keeps more of the retained earnings. Thus, any decision to invest solely on
the principle of high dividend yield would not generate abnormal returns in Indian stock
market.
152
Further, on account of lack of ability of the stocks to clear all ten screens, the
different risk-reward combinations of the criteria are estimated and tested. The results
based on market adjusted performance of the stocks revealed that in order to outperform
the market, an investor can use all the combinations safely except the combination of
stocks having market price lesser than the two-third of the net current asset value per
share along with the stocks having current ratio greater than two. Thus, it could imply
that the stocks are trading at such a low price due to the fact that the stocks have high
level of inventory which is not utilized efficiently and therefore, the earning power of
such firms is low. Further, the combination of stocks having dividend yield at least two-
third the 91 days Treasury bill yield and total debt being lesser than twice the net current
asset value per share is not recommended to invest on account of poor performance. It is
because of the reason that such firms have high dividend yield and hence lesser retained
funds. Such firms have lesser growth opportunities so they neither opt for borrowed funds
nor go for retained funds to encash future opportunities.
Along with the market adjusted performance, the abnormal returns yielded by
different risk-reward combinations on basis of asset pricing model were also assessed.
The model revealed that the abnormal returns could not be retrieved when the principle of
high dividend yield is clubbed with the risk based principles specifically, principle of low
debt in relation to book value, principle of high current ratio and the principle of growth
in earnings over a period of time. It shows that the companies having low level of
retained earnings along with low level of borrowings and huge investment blocked in
current assets could not yield extra normal returns to investors. Moreover, the firms with
high growth in earnings have high opportunity cost of dividend payout. Thus, the firms
with high growth in earnings tend to have lesser payout ratios as such firms require
retained funds for financing growth and modernization. Hence, the combination of high
earnings growth and high dividend yield, being contrary to each other, could not yield
abnormal returns in Indian stock market.
In further analysis, the entire sample of stocks has been classified on the basis of
total number of principles cleared by every stock. The results revealed that the mean
market adjusted return of the stocks that fulfilled maximum rules of the criteria were
significantly higher than the stocks that fulfilled minimum rules. The analysis also
153
showed that the mean return of the stocks becomes significantly higher than the market,
when any four or more rules of the criteria are met by the stocks. However, the stocks
meeting lesser than four rules generate lesser returns than that of market returns.
Similarly, when the abnormal return of the stocks clubbed into different categories on the
basis of number of principles met, is estimated, the same turned out to be significant in
case of stocks meeting any six or more rules in case of 12 months holding of such stocks.
Extending the holding period from 12 to 24 months, further shows that the compliance of
any four rules is sufficient to yield excess returns. The analysis also showed that the
stocks selected through Graham‟s stock selection criteria are lesser volatile than the
market. The reason behind the value premium yielded by these principles is the
behavioral or cognitive biases of the investors that yield returns to stocks having higher
intrinsic value than the market price.
Further, the predictive ability of the stock selection criteria is examined and found
that the total score formed on the basis of total number of rules followed by a particular
stock, yields positive relation with the risk adjusted returns. Moreover, when the total
score is split into ten principles and regressed against risk adjusted returns, the results
confirmed that each and every principle is not significantly affecting the stock returns.
The principles of price shrinkage, high dividend yield, high current ratio and growth in
per share earnings have lost their validity in enhancing returns in Indian stock market.
The investor, therefore, can safely invest on the basis of stocks having high earnings
yield, low price in relation to book value and net current asset value, lower debt and
stable earnings in order to attain abnormal returns in market. These principles portray
investing when price is lesser than the intrinsic value. Thus, the stock selection criteria of
Benjamin Graham which was conceived decades ago still make its place in Indian stock
market. However, each and every rule of Graham could not be used profitably in today‟s
economic environment. Thus, value investing criteria of Graham with a due consideration
to significant variables can optimize the investment decision making.
154

7 - Chapter 4

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

7 - Chapter 4

Uploaded by

Copyright:

Available Formats

CHAPTER 4

BENJAMIN GRAHAM’S STOCK SELECTION CRITERIA

4.2 TEN PRINCIPLES OF STOCK SELECTION- A THEORETICAL

Table 4.1: Results of the Significance of Market Adjusted Returns of Stocks

Table 4.2: Results of the Significance of Market Adjusted Returns of Stocks

Table 4.4: Results of the Significance of Market Adjusted Returns of Stocks

12 months holding period 24 months holding period

Table 4.5: Results of the Significance of Market Adjusted Returns of Stocks

Table 4.6: Results of the Significance of Market Adjusted Returns of Stocks

Table 4.7: Results of the Significance of Market Adjusted Returns of Stocks

Table 4.9: Results of the Significance of Market Adjusted Returns of Stocks

Table 4.11: Description of the Variables Discussed in Benjamin Graham’s

4.2.2 Testing the Difference between Fulfillment or Non Fulfillment of the

ce in Levene’s Equal Mean Levene’s

Fulfillment 604 21.218 55.228 9.398 30.068

Where, R pt is the return of portfolio p at time t, R ft is the rate of return on a risk-free

Stocks ANOVA Results Alpha Beta

4.3 TESTING THE PERFORMANCE OF DIFFERENT COMBINATIONS OF

4.3.1 Analyzing the Market Adjusted Performance of the Combinations

P2-P7 19.6662 8.7840

It is imperative to see that in case of combination of P5-P7, the returns remain

4.2.3 Analyzing the Presence of Abnormal Return Yielded by Combinations

4.4 REASONS FOR THE EXISTENCE OF VALUE PREMIUM

4.5 DETERMINING THE NUMBER OF PRINCIPLES A PARTICULAR

Total score Total no. of

The results are reported in the Table 4.20 below:

H016 : There is no significant difference in the mean market adjusted returns of

Table 4.21 shows the results of t-test.

4.5.3 Analyzing the Presence of Abnormal Returns Yielded By Different

4.6 ANALYZING THE PREDICTIVE ABILITY OF THE STOCK

4.6.1 Model Development

of 24 months) variable has been taken as dependent variable as advocated in Piotroski

4.6.1 (B) Explanatory Variables and Hypothesis Development

4.6.2 Model Estimation

(T= no. of time-series units).

Before conducting regression analysis, the association between dependent

Table 4.26: Pearson’s Product Moment Correlation Matrix

P10 4.790476 2.464487 1.94* 2.889588 1.645647 1.76*

You might also like