Value Investing with

Price-Earnings Ratio in India

Gunjan Chhaya* and Prashant Nigam**

Researchers and investment professionals have argued that value strategies based on low price relative to earnings,
dividends, book value and other fundamental measures, have outperformed the corresponding growth strategies and
the market. In this study, we endeavor to explore this premise in the Indian context by forming equity portfolios based
on price-earnings ratios and evaluate their ex post returns on both absolute and risk-adjusted measures. During the
study period, i.e., October 2000 to September 2013, we sample the market each quarter, a total of 48 iterations, and
examine the portfolio returns for holding periods up to five years. We find evidence of statistically significant value
premium in the Indian stock market.

Introduction
At its core, value investing is about investing in securities where the per-share intrinsic
value1 of the firm is greater than the price paid for acquiring it. Even as the price of a security
can be precisely quantified at any instant from the exchange ticker, determining the intrinsic
value of a firm at best delivers an imprecise valuation. Among many other dimensions, the
firms current asset valuation, broadly represented by book value, and potential to earn
profits in future, estimated with past earnings, are the two most fundamental dimensions on
which its intrinsic value is estimated. Therefore the Price-Book (P/B) ratio and the PriceEarnings (P/E) ratio, which roughly represent these dimensions, are the two most popular
measures for a firms current valuation.
For more than half a century now, researchers have investigated the effectiveness of
value strategies2 of varying complexity and sophistication, and found evidence of existence
of value premium3 based on backtesting of listed securities data. While earlier studies focused
on securities listed in US and other developed countries, there are fewer studies investigating
the existence of value premium in India.
*

Senior Consultant, Ethical Energy Petrochem Strategies Pvt. Ltd., Ahmedabad, Gujarat, India.
E-mail: gunjanchhaya@gmail.com

**

Business
Consultant,
ITC
Infotech
E-mail: prashantnigam75@gmail.com

India

Ltd.,

Bengaluru,

Karnataka,

India.

Intrinsic value is defined as that value which is justified by the facts, e.g., the assets, earnings, dividends,
definite prospects, as distinct, let us say, from market quotations established by artificial manipulation or
distorted by psychological excesses (Graham and Dodd, 1934).

Value strategies are investment strategies which choose to invest in securities based on one or more value
dimensions.

Value premium is the excess returns earned by value stocks relative to growth stocks.

2015 IUP. All Rights Reserved.

34

The IUP Journal of Applied Finance, Vol. 21, No. 2, 2015

This study endeavors to supplement the body of research in the Indian context, by focusing
on the earnings dimension and investigating investment returns, (a) over a broad liquidly
traded base-universe of 1,111 securities; (b) across 13 years of study timespan; (c) in 48
independently formed market samples; and (d) over periods ranging from one to five years of
investment holding. We do this by (i) categorizing the sampled securities on the basis of their
P/E ratios as either value stocks4 or growth stocks5; (ii) forming portfolios of value or
growth securities; (iii) measuring their ex post portfolio returns; and (iv) empirically
determining return-differentials with respect to a market benchmark. By taking a large
number of study samples, across a long study timespan, covering different economic phases
and stock market cycles, and diligent design of study methodology, we strive to avoid or
mitigate the errors in sampling such as survivorship bias6 or window dressing7 of end-ofyear corporate announcements.
We also seek to answer the fundamental questions Are these return-differentials
significant on both absolute and risk-adjusted basis? And, whether the observed value
premium is statistically significant? For this, we compare portfolio returns in absolute and
with respect to risk-adjusted measures such as Treynors ratio, Sharpes ratio, Jensens alpha,
Modiglianis M2 and Famas net-selectivity. We use ANOVA statistical procedures and
associated post-hoc tests to ascertain the statistical significance of our results. To the best of
our knowledge, no published research has examined the earnings value dimension with
similar statistical rigor and market representation. We find evidence of value premium in
the broad base of Indian securities.

Literature Review
Interest of value-based investing can be traced back to the early 1930s, when in their seminal
book Security Analysis, Graham and Dodd (1934) advocated that value strategies on a longterm basis outperform the market. These strategies focus on exploiting the margin of safety,
that is the differential between the securitys intrinsic value and its market price,
advantageously.
The academic body of research on value strategies is extensive. Nicholson (1960) was an
early proponent of superior performance of low P/E ratio strategy. In his seminal paper, Basu
(1977) studied the existence of value premium in US market and found that on an average,
both on absolute and risk-adjusted basis, the securities with low P/E ratios delivered
significantly higher returns than high P/E securities. Challenging the Efficient Market
4

Value stocks are securities with low ratio of price to earnings, book value, cash flow or other fundamental
measures of value.

Growth stocks are securities with high ratio of price to earnings, book value, cash flow or other fundamental
measures of value.

Survivorship bias is the tendency for failed or delisted companies to be excluded from performance studies
because they no longer exist.

Window dressing is the tendency to take or not take actions prior to issuing announcements in order to improve
appearance of the financial statements.

Value Investing with Price-Earnings Ratio in India

35

Hypothesis (EMH), he explored the possibility of this P/E effect explaining the violations
of Capital Asset Pricing Model (CAPM). Subsequently, Chan et al. (1991), Fama and French
(1996 and 1998), Troung (2009), Athanassakos (2009 and 2011), Gharghori et al. (2013) and
others found evidence of value premiums on different dimensions like market capitalization
(size), book value, earnings, cash flow, sales growth, debt, dividends, etc. in the international
markets.
However, the source of value premiums was keenly debated. DeBondt and Thaler (1985
and 1987) based their explanation on behavioral psychology in the form of investor
overreaction. They argued that investors tend to overreact to recent information, and price
the securities at either pessimistic or optimistic extremities. Subsequently, when these extreme
expectations prove erroneous, prices revert towards their base-rates, resulting in the observed
return anomalies. While Fama and French (1992) also found evidence of value premium,
they postulated its existence because value strategies are fundamentally risky. That is, higher
return from value strategies is essentially a compensation for assuming higher fundamental
risk. Contradicting this hypothesis, Lakonishok et al. (1994) proposed that value strategies
are not fundamentally riskier, but exploit the suboptimal investment behavior. They suggest
that factors such as judgmental errors, disregard of price for good companies, screening
constraints, investment justifiability, perceived safety, career concerns and short-time
horizons may be instrumental in causing suboptimal behavior of a typical investor.
Meanwhile, it was observed that incorporating certain risk-mitigation filters, during
formation of value portfolios, significantly improved the value premiums. Kelly et al. (2008)
found that the use of business failure prediction tests such as Altman Z-score and Castagna
and Matolcsy model yielded superior portfolio returns for low P/E ratio securities. Similarly,
Anderson and Brooks (2006) found that by smoothening the yearly variances of business
earnings, by constructing a P/E ratio based on a multi-year average instead of previous year
earnings, the value premium yield almost doubled for a potential investor. In time, as evidence
of existence of value premium was observed over greater number of markets, investors gained
confidence and started employing value strategies in their investments. Bhattacharya and
Galpin (2011) investigated 46 countries across the globe and found that value-weighted
portfolios as an investment vehicle have gained popularity over the last quarter century.
Specifically for India, however, even with a trend of increasing popularity, it is ranked among
the bottom three countries where value-weighted portfolios are the least popular.
In the Indian context, equity prices and returns have been studied from various perspectives.
Within the value investing framework, Sehgal and Tripathi (2007) evidenced the existence
of value effect in the Indian market with respect to four different value measures. Kumar et al.
(2010) applied a neural network algorithm to a collection of accounting and value ratios to
observe superior portfolio returns, while Deb (2012) studied the book value dimension on a
broad base dataset of listed companies and found prevalence of value premium on both
absolute and risk-adjusted basis. Trivedi and Behera (2012) examined and observed linkages
between macroeconomic factors and equity prices.
36

The IUP Journal of Applied Finance, Vol. 21, No. 2, 2015

Data and Methodology

This study examines the relationship between P/E ratios and the investment performance of
equity securities in India. Even as it replicates much of Basu (1977) methodology, it seeks to
strengthen the validity of the findings with the use of improved statistical procedures. For
each of the sampling periods under consideration, we ranked securities as per their P/E ratios
and formed five8 mutually exclusive portfolios based on this rank. Two additional portfolios,
representing (i) high P/E securities excluding those of loss-making firms; and (ii) market
benchmark, were also formed. Returns for these portfolios were compared for different holding
periods.
The data was drawn from the Prowess database of Centre for Monitoring Indian Economy
(CMIE). We use daily closing price, trailing 12 months earnings and dividends data of the
firms listed on Bombay Stock Exchange (BSE). The study spans a period of 13 years, with a
database of 1,111 liquid-investment-eligible securities which were traded on BSE between
October 2000 and October 2013.
Liquidity, the degree to which a security can be transacted in the market without affecting
its price, is an important criterion for investors. Many large and institutional investors
prefer, or are constrained, to avoid illiquid securities. During the study period, securities of
more than 6,000 firms were listed on BSE. However, not all these securities were liquid and
therefore were not equally investment-worthy. We therefore use S&P BSE 500 index9 (BSE
500), a group of 500 securities selected as per liquidity ranking, as a liquidity qualifier in
order to exclude illiquid securities with either low transaction volume or value. Tables 1 and
2 indicate the average daily trading volume and market capitalization for the major indices
on BSE, respectively, for the month of August 2012. It is evident that BSE 500 is representative
of the broad market, as on an aggregate basis its constituents contribute to about 85% of
average daily trading volume and 95% of market capitalization. A total of 1,111 securities
satisfied the liquidity qualifier requirements, at least at one point-of-time during the
Table 1: Average Daily Transaction Volumes of Constituents of Various Indices
of BSE for the Month of August 2012
BSE Indices
BSE SENSEX

No. of
Securities

Average Daily Turnover

(in

mn)

(in %)

Remarks

30

4,986

25

Not Considered

BSE 100

100

9,371

47

Not Considered

BSE 200

200

13,055

66

Not Considered

BSE 500

500

16,847

85

Considered

5,058

19,902

100

All BSE Securities

Not Considered

Although the formation of five portfolios is arbitrary, it ensures a large number of stocks for each portfolio.

S&P BSE 500 index inclusion criteria are, for preceding three months, (a) minimum 75% trading days;
(b) average market capitalization (75% weight); and (c) average turnover (25% weight).

Value Investing with Price-Earnings Ratio in India

37

Table 2: Average Daily Market Capitalization of Constituents

of Various Indices of BSE for the Month of August 2012
BSE Indices
BSE SENSEX

No. of
Securities

Market Capitalization
(in

bn)

(in %)

Remarks

30

29,343

48

Not Considered

BSE 100

100

44,304

72

Not Considered

BSE 200

200

52,140

85

Not Considered

BSE 500

500

58,022

95

Considered

5,058

61,288

100

All BSE Securities

Not Considered

13-year timespan of the study. We use the annualized yields of 91-day Government of India
(GoI) T-bills, available on the website of Reserve Bank of India (RBI)10, as a proxy for risk-free
asset data.
The study samples the market on the last date of each quarter. Thus, from December 31,
2000 and September 30, 2012, we form 48 sets of overlapping but independent portfolio
samples. Though the choice of sampling frequency is arbitrary, it is consistent with the
frequency of earnings updates which are announced on quarterly basis. This choice of
frequency differs from Basu (1977) who samples and forms portfolios every year. However, the
higher number of samples not only leads to greater confidence in our results as per Law of
Large Numbers (LLN)11, but also allows for greater cyclical changes in price and earnings
data.
For each sample, the sample universe consists of securities which are: (a) listed on BSE;
(b) whose past 15 months accounting data is available; and (c) are constituents of BSE 500
on the date of portfolio formation.
We computed the P/E ratio for all the securities. The numerator was defined as reported
closing price of a single common stock as on that date and the denominator as the reported
annual per-share earnings available for common stockholders. We ranked12 and sorted the
securities as per this ratio and formed five equally weighted portfolios (lowest P/E = A, B, C,
D, E = highest P/E), 100 securities each, as per this ranking. An additional portfolio, sixth13,
of 100 high P/E securities but with non-negative earnings (E*) was formed which eliminated
loss-making firms from its holdings. As aggregates, we categorize securities of low P/E portfolios
10

website: www.rbi.org

11

Borels Law of Large Numbers (LLN) states that if an experiment is repeated a large number of times, independently
under identical conditions, then the proportion of times that any specified event occurs approximately equals the
probability of the events occurrence on any particular trial; the larger the number of repetitions, the better the
approximations tend to be.

12

In practice, the reciprocal of P/E ratio, earning yield, is used in ranking the stocks. Therefore, securities with
negative earnings (loss-making) are included in the highest P/E portfolio.

13

Since including loss-making securities in highest P/E portfolio is questionable, an additional highest P/E portfolio
is formed after the loss-making securities are removed from the sample universe.

38

The IUP Journal of Applied Finance, Vol. 21, No. 2, 2015

A and B as value stocks and those of high P/E portfolios D, E and E* as growth stocks. We
required market benchmark returns for our analysis, for which a seventh portfolio (M) with
equally weighted constituents of S&P BSE 100 index14 (BSE 100) was formed as a market
proxy. These portfolios could be considered representative of the different investment
preferences which the investors may have, with respect to the earnings dimension, in their
equity investments.
Portfolio securities are assumed to be purchased on the first trading day of the subsequent
quarter, at their respective closing prices. We tracked and computed annual returns for each
of these portfolios for five subsequent holding periods (one, two, three, four and five years)
with a buy-and-hold policy, assuming equal investments across all the portfolios.
Data for both earnings and price are available at different frequencies and with varying
time lags. While the closing prices of the securities are available at the end of each trading day
without any lag, the financial data such as earnings are announced each quarter with a lag of
maximum 6015 days. We therefore consider trailing 12-month earnings ending previous quarter
for computing the P/E ratios. For example, the 27th sample period is computed on June 30,
2007, which is a non-trading day. The P/E ratios for all securities were computed with their
respective closing price as on June 29, 2007, the previous trading day, as the numerator and
12-month per-share earnings as on March 31, 2007 as the denominator. The portfolios thus
formed from ranked P/E ratios were assumed to be purchased on July 2, 2007, and sold on July
1, 2008 for one-year holding period.
The portfolio returns were computed with dividends considered to be reinvested in the
respective security on the date of their disbursal. We also adjusted individual security returns
to reflect any changes in the equity structure of the firm during the holding period. In case a
security ceased to qualify on liquidity requirements, at any point of time, it was immediately
divested and the proceeds were invested in risk-free assets for the remainder of the holding
period. Portfolio returns thus computed were annualized16 for 365-day year in order to make
them comparable with other sampling results.
For each holding period, the study evaluates portfolio performances across different
samples. We compared aggregate performances of different portfolios on both absolute and
risk-adjusted basis. Absolute returns are compared, based on summary statistics such as average
returns, standard deviation and portfolios Beta with respect to the market proxy. While we
used measures such as Treynors reward to volatility ratio; Sharpes reward to variability
ratio; and its more intuitive representation Modiglianis risk-adjusted performance (M2)
measure; Jensens alpha, a risk-adjusted absolute performance measure; and Famas net14

S&P BSE 100 index inclusion criteria are: for preceding three months, (a) minimum 95% trading days;
(b) average market capitalization (75% weight); (c) average turnover (18.75% weight); and (d) impact cost
(6.25% weight).

15

BSE listing agreement clause 41 requires all listed firms to submit their quarterly financial results within 45 days
for intermediate quarters and within 60 days for the last quarter of the financial year.

16

All returns are shown on continuously compounding basis as per the industry convention.

Value Investing with Price-Earnings Ratio in India

39

selectivity, a measure of excess return (can be negative) realized above that required to
compensate for the total risk undertaken, for risk-adjusted comparative.
The study uses statistical procedures to test the portfolio returns and infer about the
existence of statistically significant value premiums. Since we have more than two portfolios
to compare, conducting a series of t-tests could lead to higher probability of making Type I17
errors. We therefore used ANOVA18 procedure to compare multiple portfolio return-premiums
and test for statistically significant difference within them. Prior to initiating ANOVA
computations, we tested the validity of homogeneity of variance assumption with Levenes
F-test. In instances where this assumption was met, we used one-way ANOVA; while in
those where this assumption was found to be violated, Welchs ANOVA was used instead.
Finally, we undertake either TukeyHSD or Games-Howell post hoc test, as is appropriate, for
pairwise comparisons and to ascertain if the observed return-differentials are statistically
significant.

Results and Discussion

The portfolio returns were computed for all the 48 samples between January 2001 and
September 2013. We therefore have 48 samples of one-year holding returns for each of the
portfolios. Similarly, we have 44, 40, 36, and 32 samples for two, three, four, and five-year
holding returns, respectively. Table 3 shows the summary statistics of pooled portfolio
performances for all the holding periods. For all the portfolios we find that as the holding
period increases, (a) average returns improve from one to four years, after which they deteriorate
in the fifth-year; (b) volatility in returns consistently decreases, indicating increasing
confidence in the expected return values. All the portfolios, excepting portfolio D, have
their portfolio beta greater than unity, which indicates high sensitivity to market volatility
and are therefore considered to have assumed greater risk as per CAPM. On an aggregate
basis, it is observed that value stocks (portfolios A and B) consistently show superior absolute
returns than growth stocks (portfolios D, E and E*) for all the holding periods.
Table 3: Portfolio Summary Statistics
Holding
Period

Low P/E
A

High P/E
E

E*

Market
M

Mean

22.4%

22.3%

20.4%

19.6%

13.6%

14.2%

20.1%

SD

41.4%

39.2%

37.5%

35.7%

40.9%

39.1%

38.0%

Beta

1.176

1.134

1.100

1.020

1.162

1.127

1.000

1 Year

17

Comparing group means: concluding that they are different when in reality they are not, would be a false
positive or Type I error; while concluding that they are not different when in reality they are, would be a false
negative or Type II error.

18

Analysis of Variance (ANOVA) is a statistical hypothesis testing technique which is used to determine whether
there are any significant differences between means of three or more independent groups.

40

The IUP Journal of Applied Finance, Vol. 21, No. 2, 2015

Table 3 (Cont.)
Holding
Period

Low P/E
A

High P/E
E

E*

Market
M

Mean

24.0%

22.3%

22.8%

19.7%

15.8%

15.2%

21.5%

SD

25.3%

24.4%

25.8%

22.7%

25.8%

24.5%

24.0%

Beta

1.168

1.129

1.165

1.022

1.198

1.107

1.000

Mean

25.1%

24.0%

24.0%

20.5%

17.6%

16.7%

22.9%

SD

19.4%

19.9%

20.2%

16.6%

21.3%

19.7%

19.0%

Beta

1.168

1.199

1.189

0.954

1.270

1.135

1.000

Mean

26.0%

24.1%

24.0%

21.1%

18.4%

17.8%

23.2%

SD

17.1%

16.3%

16.1%

14.1%

17.9%

16.2%

16.0%

Beta

1.246

1.173

1.156

1.005

1.285

1.145

1.000

Mean

24.2%

22.7%

22.8%

19.7%

17.2%

16.5%

21.7%

SD

15.4%

14.9%

14.6%

11.4%

16.0%

13.7%

14.3%

Beta

1.250

1.197

1.193

0.917

1.303

1.074

1.000

2 Years

3 Years

4 Years

5 Years

Table 4: Portfolio Risk-Adjusted Performance Measures

Holding Period

Low P/E
A

High P/E
E

E*

Market
M

0.137

0.141

0.128

0.130

0.063

0.070

0.123

1 Year
Treynors Ratio
Sharpes Ratio

0.389

0.408

0.375

0.371

0.178

0.200

0.375

Modiglianis M2

19.1%

19.7%

18.6%

18.5%

12.2%

12.9%

18.6%

Jensens Alpha

1.6%

2.0%

0.5%

0.7%

7.0%

6.0%

t-statistic

(0.69)

(1.05)

(0.31)

(0.35)

(3.08)

(3.01)

Famas Net
Selectivity

0.6%

1.3%

0.0%

0.2%

8.1%

6.9%

Treynors Ratio

0.153

0.143

0.143

0.132

0.081

0.081

0.135

Sharpes Ratio

0.707

0.662

0.645

0.594

0.374

0.369

0.654

20.7%

19.8%

19.4%

18.4%

13.9%

13.8%

19.6%

2 Years

Modiglianis M
Jensens Alpha

2.2%

0.9%

1.0%

0.2%

6.3%

5.7%

t-statistic

(1.53)

(0.70)

(0.58)

(0.11)

(4.48)

(3.52)

Value Investing with Price-Earnings Ratio in India

41

Table 4 (Cont.)
Low P/E
A

High P/E
E

E*

Market
M

1.3%

0.2%

0.3%

1.4%

7.2%

7.0%

Treynors Ratio

0.168

0.150

0.151

0.152

0.091

0.094

0.150

Sharpes Ratio

1.014

0.901

0.890

0.869

0.543

0.543

0.932

Modiglianis M

22.4%

20.6%

20.4%

20.1%

14.8%

14.8%

21.1%

Jensens Alpha

2.2%

0.0%

0.3%

0.2%

7.3%

6.2%

t-statistic

(2.15)

(0.02)

(0.18)

(0.13)

(5.66)

(3.91)

Famas Net
Selectivity

1.6%

0.6%

0.8%

1.1%

8.3%

7.6%

Treynors Ratio

0.160

0.154

0.156

0.150

0.096

0.103

0.156

Sharpes Ratio

1.168

1.112

1.121

1.068

0.692

0.726

1.182

Modiglianis M2

21.5%

20.7%

20.8%

20.1%

15.2%

15.6%

21.6%

Jensens Alpha

0.6%

0.2%

0.1%

0.5%

7.5%

6.0%

t-statistic

(0.48)

(0.14)

(0.06)

(0.44)

(5.13)

(3.98)

Famas Net
Selectivity

0.2%

1.1%

1.0%

1.6%

8.8%

7.4%

Treynors Ratio

0.145

0.138

0.140

0.149

0.085

0.097

0.147

Sharpes Ratio

1.181

1.115

1.142

1.193

0.693

0.760

1.241

Holding Period
Famas Net
Selectivity
3 Years

4 Years

5 Years

Modiglianis M

20.1%

19.3%

19.6%

20.2%

14.3%

15.1%

20.8%

Jensens Alpha

0.2%

1.0%

0.8%

0.1%

8.0%

5.4%

t-statistic

(0.16)

(0.77)

(0.74)

(0.11)

(6.76)

(3.78)

Famas Net
Selectivity

0.9%

1.9%

1.4%

0.5%

8.8%

6.6%

Table 4 shows the portfolio performances on different risk-adjusted evaluation measures.

Treynors ratio is the beta adjusted excess portfolio returns with respect to risk-free assets.
When compared with market benchmark, it measures the portfolios risk-premium per unit
of market volatility. Sharpes ratio is similar to Treynors ratio but uses, instead, standard
deviation as a measure of risk. It therefore measures the portfolios risk-premium per unit of
variability. Modiglianis M2 is a more intuitive representation of Sharpes ratio. It adjusts the
portfolios risk-level to match the benchmarks level and measures the returns for this riskmatched portfolio. All the three measures discussed above are relative risk-adjusted measures,
that is, they require comparison with market benchmark to infer on superior or inferior
performances. Jensens alpha, on the other hand, is an absolute risk-adjusted measure. It
42

The IUP Journal of Applied Finance, Vol. 21, No. 2, 2015

measures the portfolios excess return with respect to that expected from CAPM, given the
systematic19 riskiness of the portfolio. Famas net-selectivity is an evolved absolute riskadjusted measure. It measures the portfolios excess return, after accounting for systematic
riskiness of the portfolio and unsystematic20 concentration risk assumed by the portfolio
investment strategy.
On both the relative and absolute risk-adjusted performance measures, we observe similar
evaluation inferences. We find that the value stocks show superior risk-adjusted returns for
one, two and three-year holding periods, and inferior risk-adjusted returns thereafter. While
growth stocks on the other hand show consistently inferior risk-adjusted returns for all the
holding periods. We also observe that the highest P/E portfolios E and E* show substantially
inferior performance on all the three risk-adjusted measures. This observation is strengthened
as we observe the t-statistic of Jensens alpha, where their negative alpha values for portfolios
E and E* show very high significance.
Table 5 shows the results of Levenes F-test, which tests for homogeneity of variance
within the portfolios. This test assumes that there is no difference between the group variances
as its null hypothesis; thus the rejection of its null hypothesis indicates that the assumption
of homogeneity of variance is violated. We find that for one-year holding period the F-value
is significant and the null hypothesis is rejected, i.e., at least one of the portfolios has
significantly different variance. For all other holding periods, the null hypothesis is retained
and homogeneity of variance assumption is met.
Table 5: Equality of Variance
Holding Period (Between)
DF

(Within)
DF

F-Value

Significance
p-Value

Remarks

1 Year

282

2.299

0.045*

Null Rejected

2 Years

258

0.342

0.887

Null Retained

3 Years

234

0.945

0.453

Null Retained

4 Years

210

0.934

0.460

Null Retained

5 Years

186

1.229

0.297

Null Retained

Note: (a) Level of significance set a priori at = 0.05 to test the assumption of homogeneity of variance.
(b) * indicates significance at 0.05% level.

Homogeneity of variance is the key underlying assumption for one-way ANOVA.

Therefore, for one-year holding period, where this assumption is found to be violated, we
apply Welchs ANOVA, which does not assume equal variance between groups, and adjust
the F-statistic accordingly. For all the other holding periods, as the underlying assumption is
met, we undertake one-way ANOVA. Table 6 shows ANOVA results for different holding
19

Systematic risk or market risk is the uncertainty inherent in the entire market (or segment). Investors cannot
reduce systematic risk by diversifying within the same asset class (or segment).

20

Unsystematic risk or specific risk is the uncertainty assumed by concentrating investments. Invest or can
reduce unsystematic risk by diversifying within the same asset class (or segment).

Value Investing with Price-Earnings Ratio in India

43

Table 6: Analysis of Variance

Holding Period (Between)
DF

(Within)
DF

F-Value

Significance
p-Value

Remarks

1 Year $

131.16

3.58

0.005**

Null Rejected

2 Years

258.00

7.76

0.000***

Null Rejected

3 Years

234.00

12.41

0.000***

Null Rejected

4 Years

210.00

12.36

0.000***

Null Rejected

5 Years

186.00

14.41

0.000***

Null Rejected

Note:

For 1-year holding period, we adjust F-statistic with Welchs procedure as the equal variance assumption
was violated.
** and *** indicate significance at 0.01% and 0.001% levels, respectively.

periods. The null hypothesis in ANOVA is that the population means from which the
samples are selected are equal, and the alternative hypothesis is that at least one of group
means significantly differs from other portfolio means. Since we see that for all the holding
periods the F-statistic is significant, we reject the null hypothesis and conclude that there is
at least one portfolio whose returns are significantly different from other portfolio returns.
Table 7 shows the results of the post hoc tests. These tests undertake multiple pairwise
comparisons to find the statistical significance of the differences between portfolio means.
As is appropriate, we use Games-Howell test for the one-year holding comparisons, and
TukeyHSD test for all the other holding periods. We find that value stocks consistently show
superior performance as compared to growth stocks, indicating statistically significant value
premium for all the holding periods.
Table 7: Post Hoc Multiple Comparison Tests
Holding Period
$

Multiple Pairwise Comparison

1 Year

Games-Howell Test

0.16

2.07

2.89

8.86^

8.29^

1.90

2.73

8.70*

8.12*

0.83

6.79

6.22

5.97

5.40

C
D

E
2 Years

TukeyHSD Test

0.57
1.73

1.20

4.36

8.21***

8.84***

0.53

2.63

6.48**

7.11**

3.16

7.01**

7.64**

3.85

4.47

C
D
E
3 Years

TukeyHSD Test

44

E*

0.63
1.70

1.69

5.17*

8.07***

8.96***

0.02

3.47

6.37***

7.25***

The IUP Journal of Applied Finance, Vol. 21, No. 2, 2015

Table 7 (Cont.)
Holding Period

Multiple Pairwise Comparison

B

E*

3.48

6.38***

7.27***

2.90

3.79

D
E
4 Years

TukeyHSD Test

0.89
1.84

1.92

4.87**

7.55***

8.19***

0.08

3.03

5.71***

6.35***

2.95

5.63***

6.27***

2.68

3.33

C
D
E
5 Years

TukeyHSD Test

0.64
1.56

C
D
E
Note:

1.44
0.12

4.51**

7.04***

7.74***

2.96

5.48***

6.18***

3.08

5.60***

6.30***

2.52

3.22^
0.70

For 1-year holding period we use Games-Howell Test, which is an appropriate post hoc test for Welchs
ANOVA.
*** , ** and * indicate significance at 0.001%, 0.01% and 0.05% levels, respectively; and ^ indicates
significance at 0.1%.

Conclusion
This study examines the performance of low P/E stocks over high P/E stocks in the Indian
stock market from the year 2000 to 2013. We find that, for all buy-and-hold horizons, the
value stocks outperform and the growth stocks underperform the market benchmark for
a major part of the study period. This performance differential is observed on both absolute as
well as risk-adjusted performance measures such as Treynors ratio, Sharpes ratio, Modiglianis
M2, Jensens alpha and Famas net-selectivity. On an aggregate basis, evidence of consistent
value premium was observed with respect to earnings dimension. This was statistically
validated by ANOVA. Finally, post hoc multiple pairwise comparison tests indicate that the
performance differential is more pronounced between the extreme value and growth portfolio
pairs. Thus, the study findings suggest that risk does not adequately explain the existence of
value premium in India. The study findings are relevant not only for individual investors
interested in investing in India, but also for institutional investors who manage thematic
investments and are interested in confirming the pervasiveness of value premiums in India.
Limitations: We used BSE 500 as a liquidity qualifier which limited our investment universe
to 500 stocks only at any point of time. Though liquidity represents a practical constraint for
many large investors, it does exclude from the study a majority of the listed but illiquid stocks
which may be of interest to an individual investor. The stock returns computation does not
Value Investing with Price-Earnings Ratio in India

45

consider transaction and holding costs like brokerage charges, transaction taxes, account
maintenance charges, etc.
Also, at a conceptual level, under certain circumstances, value stocks may indeed carry
greater investment risk than growth stocks. Special situations in the firms business like
severe financial distress or bankruptcy, industrial supply overcapacity or demand destruction,
sudden policy changes like import or export restrictions, geopolitical threats, regulatory or
legislative impositions, etc. may lead to extreme valuations and low P/E ratio. Some of these
special situations have quantitative proxies, but others remain a matter of subjective
qualitative assessment. Value investors typically do incorporate such business dimensions in
their investment framework. Even as we do not factor them in this study, we consider it a
topic for interesting future research.
The study evaluates performance differential with respect to only one accounting
dimension: earnings. Fundamental investing may include scrutiny of stocks on other value
dimensions like book value, size, cash flow, dividend, etc.; business dimensions like equitydebt structure, growth, capital efficiency, sales, profitability, etc.; and qualitative dimensions
like industry characteristics, competition, product life cycle, regulations, demand-supply,
etc. In addition, macroeconomic factors may also be influential in determining the extent of
success of an investment strategy. We would like to augment our study by evaluating equity
returns with respect to a combination of value and business dimensions.
Acknowledgment: The authors gratefully acknowledge the assistance extended by Entrepreneur Development
Institute of India (EDII) for providing access to its library resources. The authors also thank Dr. Anoop Singh
of IME Department, IIT Kanpur, for providing guidance for the study.

References
1. Anderson Keith and Brooks Chris (2006), The Long-Term Price-Earnings Ratio, Journal
of Business Finance and Accounting, Vol. 33, Nos. 7-8, pp. 1063-1086.
2. Athanassakos George (2009), Value versus Growth Stock Returns and the Value Premium:
The Canadian Experience 19852005, Canadian Journal of Administrative Sciences,
Vol. 26, No. 2, pp. 109-121.
3. Athanassakos George (2011), The Performance, Pervasiveness and Determinants of Value
Premium in Different US Exchanges: 1985-2006, Journal of Investment Management,
Vol. 9, No. 3, pp. 33-73.
4. Basu Sanjoy (1977), Investment Performance of Common Stocks in Relation to Their
Price-Earnings Ratios: A Test of Efficient Market Hypothesis, Journal of Finance, Vol. 32,
No. 3, pp. 663-682.
5. Bhattacharya Utpal and Galpin Neal (2011), The Global Rise of the Value-Weighted
Portfolio, Journal of Financial and Quantitative Analysis, Vol. 46, No. 3. pp. 737-756.
46

The IUP Journal of Applied Finance, Vol. 21, No. 2, 2015

6. Chan Louis, Hamao Yasushi and Lakonishok Josef (1991), Fundamentals and Stock
Returns in Japan, Journal of Finance, Vol. 46, No. 5, pp. 1739-1764.
7. Deb Soumya Guha (2012), Value Versus Growth: Evidence from India, IUP Journal of
Applied Finance, Vol. 18, No. 2, pp. 48-62.
8. DeBondt Werner and Thaler Richard (1985), Does the Stock Market Overreact, Journal
of Finance, Vol. 40, No. 3, pp. 793-805.
9. DeBondt Werner and Thaler Richard (1987), Further Evidence on Investor Overreaction
and Stock Market Seasonality, Journal of Finance, Vol. 42, No. 3, pp. 557-581.
10. Fama Eugene (1972), Components of Investment Performance, Journal of Finance,
Vol. 27, No. 3, pp. 551-567.
11. Fama Eugene and French Kenneth (1992), The Cross-Section of Expected Stock
Returns, Journal of Finance, Vol. 47, No. 2, pp. 427-465.
12. Fama Eugene and French Kenneth (1996), Multifactor Explanation of Asset Pricing
Anomalies, Journal of Finance, Vol. 51, No. 1, pp. 55-84.
13. Fama Eugene and French Kenneth (1998), Value versus Growth: The International
Evidence, Journal of Finance, Vol. 53, No. 6, pp. 1975-1999.
14. Gharghori Philip, Stryjkowski Sebastian and Veeraraghavan Madhu (2013), Value versus
Growth: Australian Evidence, Accounting and Finance, Vol. 53, No. 2, pp. 393-417.
15. Graham Benjamin and Dodd David (1934), Security Analysis, McGraw-Hill of New York,
New York.
16. Jensen Michael (1968), The Performance of Mutual Funds in the Period 1954-1964,
Journal of Finance, Vol. 23, No. 2, pp. 389-416.
17. Kelly Simone, McClean Jenna and McNamara Ray (2008), The Low P/E Effect and
Abnormal Returns for Australian Industrial Firms, 21st Australasian Finance and
Banking Conference, Sydney.
18. Kumar Krishna M S, Subramanian S and Rao U S (2010), Enhancing Stock Selection in
Indian Stock Market Using Value Investment Criteria: An Application of Artificial
Neural Networks, IUP Journal of Accounting Research and Audit Practices, Vol. 9, No. 4,
pp. 54-67.
19. Lakonishok Josef, Shleifer Andrei and Vishny Robert (1994), Contrarian Investment,
Extrapolation and Risk, Journal of Finance, Vol. 49, No. 5, pp. 1541-1578.
20. Modigliani Franco and Modigliani Leah (1977), Risk-Adjusted Performance, Journal
of Portfolio Management, Vol. 23, No. 2, pp. 45-54.
21. Nicholson Francis (1960), Price-Earnings Ratios, Financial Analysts Journal, Vol. 16,
No. 4, pp. 43-45.
Value Investing with Price-Earnings Ratio in India

47

22. Sehgal Sanjay and Tripathi Vanita (2007), Value Effect in Indian Stock Market, IUP
Journal of Applied Finance, Vol. 13, No. 1, pp. 23-36.
23. Sharpe William (1966), Mutual Fund Performance, Journal of Business, Vol. 39, No. 1
(Supplement), pp. 119-138.
24. Treynor Jack (1965), How to Rate Management Investment Funds, Harvard Business
Review, Vol. 43, Nos. 1-2, pp. 63-75.
25. Trivedi Pushpa and Behera Samir Ranjan (2012), Macroeconomic Fundamentals as
Determinants of Equity Prices: An Empirical Analysis for India, IUP Journal of Applied
Finance, Vol. 18, No. 3, pp. 5-30.
26. Truong Cameron (2009), Value Investing Using Price Earnings Ratio in New Zealand,
University of Auckland Business Review, Vol. 11, No. 1, pp. 1-7.

Reference # 01J-2015-04-03-01

48

The IUP Journal of Applied Finance, Vol. 21, No. 2, 2015

Copyright of IUP Journal of Applied Finance is the property of IUP Publications and its
content may not be copied or emailed to multiple sites or posted to a listserv without the
copyright holder's express written permission. However, users may print, download, or email
articles for individual use.