Chapter-10 Efficient Market Hypothesis

1. Introduction to Efficient Market Hypothesis (EMH)

Analysts and Investors use the fundamental analysis and technical analysis in order to analyze an
investment option and take investment decisions. Fundamental analysis is used to determine if the
stocks are properly priced or not and technical analysis is used to find out the price and volume
patterns in the past and predict the likely future price trends. Besides these analyses, there are few
more other tools that form the basis for investment decisions.
Efficient Market Hypothesis
The EMH opines that the capital markets are efficient and all the information available in the market
is reflected in the stock prices. The market adjusts to the new information and therefore the returns to
the investor will vary depending on the efficiency of the market. According to EMH, the participants
in the market are rational and do not affect the movement in the stock prices.
Forms of EMH
 Weak Form of EMH: According to the weak form, the prices reflect the historical
information in the market and therefore an investor cannot earn more returns based on his
predictions about future prices depending on the past price trends. This form is also known
as the "Random Walk Theory".
 Semi-strong Form of EMH: According to this form, the prices in the market reflect all the
information that is made public like the stock splits, dividends etc. Though, it may take
some time to absorb the new information, it will be done in a very short-term. Therefore,
the analyst cannot use the new information to make excess returns.
 Strong Form of EMH: According to this form, the prices reflect not only the publicly
available information but also private information which is available to selected groups like
management, financiers, etc., about the company.
- Near-strong Form: According to it, the prices reflect all the publicly available information.
- Super-strong Form: According to it, the prices contain an adjustment for the confidential
information which is reflected in the prices.
Though there are transaction costs , information costs etc., that might make the market inefficient,
these are not present in a market to such an extent that a trading strategy might be designed based on
them to make profits.
3. Empirical Evidence of Tests of EMH
Empirical Evidence
According to EMH, the price changes are independent of each other and investors in the market are
rational. However, the reality is investors and security analysts are quite irrational and in the Indian
context the prices are affected mostly by the investors' sentiments.
Investors in general hold only those stocks that are popular and those that are performing well. They
do not try new asset classes like commodities, real estate, international stocks.
They do mental calculations which often lead to confusion and use more of historical data to predict
the future. They follow a herd tendency and are over confident about their strategies. Empirical
evidence also shows that the investors do not form an efficient portfolio of securities. Security
analysts on the other hand have a biased approach towards investments. They are optimistic as
compared to the investors' pessimism. The new issues (IPO companies) have more anomalies than the
old ones due to their optimistic bias about new firms. Investors in general are more gullible than
analysts.
Tests of Weak Form
 Serial Correlation Test: The weak form or the random walk theory suggests that prices in
different periods are independent.
 Filter Tests: Using the filters, the returns from buy and hold strategy and returns from filter rules
were compared. The test done by one Mr. Rao showed that returns under the filter rules were
lesser than the other strategy (transaction costs excluded).
Tests of Semi-strong Form
These tests verify whether the prices absorb the publicly available information in the market and still
the investor can use this information and make profits.
4. Tests of EMH
Tests of Weak Form of EMH
Test of Weak Form of EMH test whether the price in different periods are related or not and also
whether or not an investor can make profit using the historical information.
Auto-correlation Test
Few stocks are chosen and the changes in their prices in two periods are studied for which correlation
analysis is conducted. The prices are said to be independent if the correlation between these
changes is zero or near zero.
r2 = a ∑ Y + b ∑ XY-n. Y 2
∑ Y2-nY 2
∑ XY −n XY
Where, 'b' is given by and 'a' by Y-bX .
∑ X 2 n X ❑2
The magnitude of correlation is r.
Example 1
Mr. Praneet Sharma feels Indian stock market during January 2005, has exhibited weak form of
market efficiency. In order to test the validity of his impression, he collected following data relating to
movement of Nifty for the 20 trading days in January 2005:
Date Open High Low Close
Ol-Jan-2005 1880.35 1917.05 1880.35 1912.25
02-Jan-2005 1912.25 1951.70 1911.05 1946.05
05-Jan-2005 1946.30 1969.20 1930.75 1955.00
06-Jan-2005 1955.10 1979.05 1908.75 1926.70
07-Jan-2005 1927.95 1930.95 1888.10 1916.75
08-Jan-2005 1918.10 1973.45 1918.10 1968.55
09-Jan-2005 1969.00 2014.65 1957.45 1971.90
12-Jan-2005 1972.00 1980.55 1936.75 1945.60
13-Jan-2005 1944.70 1967.85 1926.10 1963.60
14-Jan-2005 1987.40 1995.20 1970.10 1982.15

15-Jan-2005 1983.20 2000.30 1933.25 1944.45

16-Jan-2005 1944.15 1953.05 1887.10 1900.65
19-Jan-2005 1901.90 1943.10 1874.95 1935.35
20-Jan-2005 1928.80 1957.65 1876.85 1893.25
21-Jan-2005 1895.45 1899.55 1811.35 1824.60
22-Jan-2005 1824.70 1854.55 1756.25 1770.50
23-Jan-2005 1771.10 1858.50 1771.10 1847.55
27-Jan-2005 1847.90 1911.30 1844.65 1904.70
 28-Jan-2005 1903.90 1918.45 1846.35 1863.10
29-Jan-2005 1863.00 1883.10 1827.25 1843.60

Test the weak form of efficient market hypothesis using the auto-correlation test, taking a time lag of
10 days.

Solution:-
Period I Nifty closing Change (X) Period II Nifty Change (Y)
closing
1 1912.25 - 11 1944.45 -
2 1946.05 33.8 12 1900.65 -43.8
3 1955.00 8.95 13 1935.35 34.7
4 1926.70 -28.3 14 1893.25 -42.1
5 1916.75 -9.95 15 1824.60 -68.65
6 1968.55 51.8 16 1770.50 -54.1
7 1971.90 3.35 17 1847.55 77.05
8 1945.60 -26.3 18 1904.70 57.15
9 1963.60 18 19 1863.10 -41.6
10 1982.15 18.55 20 1843.60 -19.5

X Y X2 Y2 XY
33.80 -43.80 1142.44 1918.44 -1480.44
8.95 34.70 80.1025 1204.09 310.565
-28.30 -42.10 800.89 1772.41 1191.43
-9.95 -68.65 99.0025 4712.8225 683.0675
51.80 -54.10 2683.24 2926.81 -2802.38
3.35 77.05 11.2225 5936.7025 258.1175
-26.30 57.15 691.69 3266.1225 -1503.045
18.00 ^1.60 324 1730.56 -748.8
18.55 -19.50 344.103 380.25 -361.725
69.90 -100.90 IX2 = 6176.6 EY2 = 23848.2 EXY =-4453.2

X = 69.9/9 =7.767 (X)2 =60.33

Y = -100.9/9 =-11.21 (Y ) 2 =125.66
r2 = a ∑ Y + b ∑ XY-n. Y 2
∑ Y2-nY 2
a = Y - bX
∑ XY −n XY
b= 2 2
∑X n X❑
−4453.2 – 9 ( 7.767 x−11.217 ) −4453.2+783.61
b= =
6176.6−( 9 x 60.33 ) 6176.6−542.97
3669.6
b= = -6.154
5633.63
a = -11.21-(-0.651) x (7.767) = -6.154
r2 =(−6.154 X−−100.9 ) + (−6.154 x−4453.2 )−¿ ¿ = 0.1052
 r = √ 0.1052 = -0.3243
Hence, there is small degree of correlation between the returns of two periods and we can
conclude the market does not exhibit weak form of efficiency.
Runs Test
The randomness of the prices is tested by using runs test. Here, the stock prices for a certain time
period are taken. As compared to the preceding price, an increase is indicated by '+', a decrease is
indicated by '-'and when there is no change, it is indicated as zero. A change in the sign indicates the
beginning of a new run. The number of positives is denoted as 'n,' and a number of negatives is
denoted by 'n2'and the number of runs is denoted by 'r\ The mean and standard error of the runs is
calculated and checked to see if the runs fall within the upper or lower limits at a given level of
significance.

Tests of Semi-strong Form of EMH

To determine the validity of this form that the investors will not profit by using publicly available
information the following tests are used.
Residual Analysis
Using a regression equation of the security returns with the market index, the residual returns are
calculated. The error term represents the residual return and for an efficient market it should be near
or equal to zero.

The error term should be zero as the deviations in the security's returns from the estimating line cancel
each other. Therefore, the normal return is

The residual returns therefore is, eit = rlt - (a, + (3; rmt) [1-2]
Example 2
Mr. Sivam Gupta is a young analyst who is encouraged by the performance of Optima Labs. He
discovered that announcement regarding approval of a patent right induced a sharp interest in the
scrip. To check the market efficiency in semi-strong form in the above case, Mr. Gupta collected the
following relevant information:
Month Closing Price of Closing Value of
Optima Labs (Rs.) Market Index
November, 2003 125.25 1015
December, 2004 135.25 1085
January, 2005 145.25 1095
February, 2005 156.75 1110
March, 2005 136.75 1090
April, 2005 175.25 1150
May, 2005 205.25 1175
June, 2005 225.25 1215
July, 2005 215.25 1285
August, 2005 245.25 1345

Further using the data for 3 previous years, the characteristic line arrived by Mr. Gupta is
rs,t = 4.52 + 0.78rm,t
Where,
rm,t is % monthly return on market in any month t
rS,t is % monthly return on optima stock in the same month t.
Conduct the residual analysis to test semi-strong form of market efficiency.
Solution
Months (1) Actual Return on Expected Return
Return Index on Reddy's Above Normal
on Stock (%) (III) stock using Return % (V) =
(%) characteristic line (II) - (IV)
(II) (IV)
November, -
2004
December, 7.984 6.897 9.899 -1.915
2004
January, 7.394 0.922 5.239 2.155
2005
February, 7.917 1.37 5.588 2.329
2005 2005 -12.8
March, -1.8 3.115 -15.87
April, 2005 28.15 5.505 8.814 19.34
May, 2005 17.12 2.174 6.216 10.9
June,2005 9.744 3.404 7.175 2.569
July, 2005 -4.44 5.761 9.014 -13.45
August, 2005 13.94 4.669 8.162 5.775
Z Above normal 11.83
return =
From the above computation, we observe that sum of abnormal return is not close to zero. Therefore,
we conclude that market is not efficient in semi-strong form.
Adjustment for Market Effects
The security's returns will have to be adjusted for the rates of return of the overall market to test the
alternative EMH. As such, the abnormal return is calculated by subtracting the market return from the
security's return.
 ARa = Rit - Rmt
However, as some stocks are more volatile and some are less, the relationship between the stock's
relationship with the market and the expected rate of return for the stock based on the market rate of
return has to be determined. Then the abnormal rate of return can be calculated as:
However, as some stocks are more volatile and some are less, the relationship between the stock's
relationship with the market and the expected rate of return for the stock based on the market rate of
return has to be determined. Then the abnormal rate of return can be calculated as:
ARa = Rit-E(Rit)
Where,
E(Rit) = Expected rate of return for stock 'i' during period ‘t' based on the market rate of return and the
stock's normal relationship with the market.
Two-sets of Semi-strong EMH
In one test, the analysts will look for information like P/E, dividend yield, etc., to predict cross-
sectional distribution of risk adjusted rates of return. Whereas, in the second test, the abnormal rates
of return are examined, to know if any investor made above average risk adjusted rates of return
subsequent to a significant announcement.
Results of Return Prediction Studies
The time series tests are done to find out whether any publicly available information will provide
higher return estimates in the short-run or the long-run. The studies based on returns showed that it is
easy to predict for the long-run than the short-run. The studies used variables like (i) the difference
between yields on lower rated bonds and higher rated bonds, the risk premium, (ii) the difference
between yield on long-term higher rated bonds and yield on short-term bonds referred to as the term
structure of horizon spread. The studies further showed that the predictive power of several economic
variables related to stock returns changed with time and volatility in returns. The economic indicators
can be used to predict the returns in volatile markets rather than in calm markets.
Quarterly Earnings Studies
These studies predict the future stock returns based on the quarterly earnings reports. These studies
were based on certain categories to determine the deviations from expectations. That is (i) any
deviation from expectations, (ii) a deviation plus or minus 20 percent, and (iii) a deviation of at least
40 percent. The studies showed that favorable information in quarterly earnings did not immediately
affect the stock prices, however surprises in quarterly earnings reflected in the stock prices. Further,
the returns were considerably optimistic after the announcement conflicting market efficiency. The
recent studies use Standardized Unexpected Earnings (SUE) to find the differences between actual
and unexpected earnings for the quarter.

SUE = ________________Reported EPS, - Predicted EPSt______________

Standard Error of Estimate for the Estimating Regression Equation
The standard deviation measures the standard error of a statistic. The studies done to find the affect of
earnings surprise showed that over 80% of the stock prices moved smoothly following
announcements about unexpected earnings. As such the unexpected earnings can be used to predict
returns on individual stocks unlike the fast absorbance of information as suggested by semi-strong
form of EMH.

Calendar Studies
Studies were conducted to check if returns could be predicted for any calendar year and further was
there any pattern that would help investors predict the returns in advance.
January Anomaly: The studies showed that investors adopted selling the shares in the year end that
did not perform well during the year to acquire either same shares or other attractive shares. As such,
there was downward pressure in the last two months and upward pressure in January. Further, there
was abnormally high volume for stocks that underperformed and low volume for shares that
outperformed the previous year. A non-liner relationship was observed during January between
dividend yields and stock return. More than 50% of the January effect was more focused in the first
week and particularly on the first day of the year. Further, the trading volume and dividend yields
were affected by seasonal impacts.
Other Calendar Effects: Studies were conducted to examine the weekend and start of the week
affects. Results showed that the mean return was negative on Mondays for individual stock and T-
bills. An analysis of NYSE transactions indicated that the negative Monday effect occurred before the
market opened and for smaller firms it occurred during the day on Monday.
Predicting Cross-Sectional Returns
On the assumption that the markets are efficient, the securities should lie on the SML. Further, in an
efficient market the prices absorb all the information available in the market and all the securities
should have the same risk-adjusted returns. The cross-sectional returns can be forecasted using the
size or quality as tools to rank securities in terms of risk-adjusted returns.
EMH can be tested using the relationship between the historical P/E for stocks and their returns. It
was observed that firms with low P/E outperformed in the long run than the stocks with high P/E.
Investors can use the publicly available information about P/E to predict returns on the stocks. When
P/Es were adjusted for firm size, industry effects and infrequent trading of the risk-adjusted returns of
stock with low P/E were more than for those with high P/E.
Studies showed an inverse relationship between PEG ratios and subsequent returns. It was found that
the tests used to measure the PEG ratios did not support EMH.
An examination of the impact on risk adjusted returns showed that small firms gave high risk-adjusted
returns than larger firms. The influence of firm size on predicting future returns however remains an
anomaly.
Event Studies
These test how abnormal rates of return are affected by economic information. Evidence shows that
stock splits, declaration of dividends, announcement of other economic information pertaining to the
world or the country had a considerable effect on the changes in the stock prices. The results suggest
that a stock-spilt did not result in an increase in the prices of the stocks thus supporting the semi-
strong form of EMH. As far as IPOs are concerned, the price adjusts in a day but the investor cannot
make abnormal profits. However, an investor could make good profits between the period of
announcement and period of listing, thus violating the semi-strong form of EMH. The studies further
showed that the prices adjusted to the economic information even before the market opened.
Example 3
During the year 2004-05, three companies Polar Software Ltd., Sonata Airways and Time Auto Ltd.,
announced higher dividends on December 31, 2004. A financial analyst working in a brokerage firm
wanted to test the consistency of the semi-strong form of market efficiency. He estimated the
characteristic lines for a period of 4 years on a monthly basis upto September 30, 2004. The
relationship between the returns on these three companies and the market index are represented by the
following equations:
rPt = 1.50% + 0.75rmt
rs,t = 1.26%+1.15 rmt
rt,t = 1.98%+ 1.35 rmt
Where rP,t rS,t and rw are the returns of Polar Software, Sonata Airways and Time Auto during period t
and rm,t is return of the market index during the same period. The following data pertains to the
returns of the companies and market for the period of 3 months before and 3 months after the
dividend was declared.
Period (Months) Actual Return (%) Market Return (%)
r
Pt
r,
st rt,t rmt
Sep 30, 2004 9.75 12.45 14.58 9.85
Oct 31, 2004 9.85 12.35 14.85 9.95
Nov 30, 2004 10.25 12.85 15.35 10.05
Dec 31, 2004 10.45 13.45 15.78 10.25
 Jan 31, 2005 10.75 13.38 16.15 10.45
Feb 29, 2005 11.25 14.25 17.35 11.25
Mar 31,2005 10.85 14.15 17.95 11.45

Using the event studies approach verify the validity of semi-strong form of market efficiency in the
Indian stock market.

Solution
First, we should find out abnormal return by deducting the actual return from the expected return.
Polar, Software Limited
Period Actual Market Expected Abnormal
Return Return Return Return
(rPt) (1) (ftm) (2) (%) (1.50 + 0.75 rmt) (1-2)

3 9.75 9.85 8.8875 0.8625

2 9.85 9.95 8.9625 0.8875
1 10.25 10.05 9.0375 1.2125
0 10.45 10.25 9.1875 1.2625
1 10.75 10.45 9.3375 1.4125
2 11.25 11.25 9.9375 1.3125
3 10.85 11.45 10.088 0.7625

Sonata Airways
Period Actual Market Expected Abnormal
Return Return Return Return
(rst) (rmt) (%) (1.26 +1.15 (rmt)

3 12.45 9.85 12.5875 -0.1375

2 12.35 9.95 12.7025 -0.3525
1 12.85 10.05 12.8175 0.0325
0 13.45 10.25 13.0475 0.4025
1 13.38 10.45 13.2775 0.1025
2 14.25 11.25 14.1975 0.0525
3 14.15 11.45 14.4275 -0.2775

Time Auto
Actual Market Expected Abnormal
Period
Return Return Return Return
(rT,t) (rrat) (1.98 +1.35 rrat)

3 14.58 9.85 15.2775 0.6975

2 14.85 9.95 15.4125 0.5625
1 15.35 10.05 15.5475 0.1975
0 15.78 10.25 15.8175 0.0375
1 16.15 10.45 16.0875 -0.0625
2 17.35 11.25 17.1675 -0.1825
3 17.95 11.45 17.4375 -0.5125

We will now estimate the average abnormal return to each of the months before and after the dividend was
announced.
Third month before the announcement of dividend

AAR(-3) = - (0.8625-0.1375 + 0.6975)= 1.4225

Second month before the announcement of dividend

AAR(-2) = - (0.8875- 0.3525 + 0.5625) = 1.0975

First month before the announcement of dividend

AAR(-1) = - (1.2125+0.0325+ 0.1975) =1.4425

Month during which the dividend was announced

AAR(0) = - (1.2625+ 0.4025+0.0375) =1.7025

First month after the announcement of dividend

AAR(1) = - (1.4125 + 0.1025 - 0.0625) = 1.4525

Second month after the announcement of dividend

AAR(2)= - (1.3125+0.0525-0.1825) =1.1825

Third month after the announcement of dividend

AAR (3) = - (0.7625-0.2775-0.5125) =-0.0275

Now, we will compute the Cumulative Average Abnormal returns for the period of three months
before and after the announcement of dividend.

CAAR = (1.4225 + 1.0975 + 1.4425 + 1.7025 + 1.4525 + 1.1825 -0.0275) = 8.2725%.

As the value of CAAR is not close to zero, we conclude that market is not efficient in the semi-strong
form.

Chaos Theory

Chaos theory consists of chaos models which are deterministic mathematical models. The
observations resulting from chaotic models have the following properties:

They appear random (even when subject to sophisticated statistical analysis).

 No pattern repeats itself. Thus even with infinite data, no two patterns will ever be exactly
alike.
 The time series is subject to sharp and substantial breaks in the usual pattern or trend, which
are completely undetermined by what went before. Thus, a series of observations with a high
degree of volatility and an upward trend can be suddenly replaced by a flat downward trend.
 The qualitative behavior of a chaotic time series is subject to complete upheaval in response
to the most microscopic change in the values of the underlying parameters.

Tests of Strong-form of EMH

The validity of this form that publicly and privately available information is reflected in the stock
prices can be established from the following:

Trading by Exchange Officials

A study conducted in the US regarding the profits made by the specialists or the top officials of the
stock exchange indicated that they made abnormal profits by using the information they are privy to.
There by violating the super-strong form of EMH.

Trading by Mutual Fund Managers

The mutual fund managers use the information available publicly but not accessible to the common
man. However, studies show that the mutual fund managers have not made considerable returns
violating the near-strong form of EMH.

Insider Trading

If stock markets are to run efficiently then, it is a prerequisite to have a strong regulatory body that
tackles insider trading. Insider trading is prohibited in India and SEBI has laid down rules regarding
insider trading. Any person who is privy to certain information should not either directly or indirectly
advice investors based on privy information. In case of a receipt of allegation about insider trading
investigations are conducted and if found guilty the insiders are heavily penalized.

Annexure
Rational investor is the one who maximizes the utility or want higher return for same level of risk or
lower risk for same level of return.