第七章 效率市场主要内容
效率市场的定义
效率市场的实证结果
证券市场中的几种反常现象
行为金融
Maurice Kendall,The analysis of
economic time series,Part I,Prices,
Journal of the Royal statistical society
96,1953.
He could identify no predictable patterns
in stock prices.
1,Why are not stock prices
predictable?
例子：一种模型预测股票价格三天后将从 20元 /股涨至 23元 /股
The forecast of a future price increase
will lead instead to an immediate price
increase
the stock price will immediately reflect
the good news implicit in the model’s
forecast.
例子,2001年 10月 21日晚，证监会宣布暂停国有股减持
例子,2002年 4月 5日宣布，从 5月 1日起交易佣金使用 0.3%上限向下浮动制。
例子：中国人民银行决定，从 2004年 10月 29日起上调金融机构存贷款基准利率并放宽人民币贷款利率浮动区间和允许人民币存款利率下浮。
金融机构一年期存款基准利率上调 0.27个百分点，由现行的 1.98%提高到 2.25%，一年期贷款基准利率上调 0.27个百分点，由现行的
5.31%提高到 5.58%。
由于存在聪明的投资者的原因，任何能够用来对股票价格作预测的信息已经反映在股票的价格中。
任何新信息，如果是可以预测的，则已经反映在价格中；如果是不可预测的，
则导致的股票价格变动也是不可预测的，
即随机的信息导致随机的股票价格变化。
2,股票价格的随机游走 (random walk)
股票价格的变动服从随机游走的形式。
随机游走的形式并不是说明市场是非理性的，而恰恰表明这是投资者真相寻求相关信息，以使得自己在别的投资者获得这种信息之前买或者卖股票而获得利润的结果。
反之，如果股票价格不是随机变动而是可以预测的，则表明所有的相关信息并没有完全反映在价格中，这表明市场是非有效的。
股票价格反映了所有可得信息称为 有效性市场假设 。
有效资本市场指的是现时市场价格能够反映可得信息的资本市场，在这个市场中，不存在利用可得信息获得超额利润的机会。
资本市场的有效性这一概念对于金融投资者而言具有非常重要的意义。因为市场的有效性消除了许多可以提高收益的策略。
Market efficiency means
Prices are correct
They fully reflect all available information
People use all available information in forming expectations
about future cash flows.
The discount rate is right for the riskness of the cash flows.
Prices react to new information quickly and to the right extent
There is no free lunch
The only way you can get higher returns is by taking on more risk
There is no information out there that can be used to construct
strategies that earn returns higher than required for their risk.
When we say ‘prices are correct’,we are implicitly statement what
‘correct’ is (i.e.,we are assuming an asset pricing model)
1 1i i
ittt
r
dEP
3,有效资本市场的描述
例子,F-stop Camera Corporation is
attempting to develop a camera that
will double the speed of the auto--
focusing system now available,
在特定的价格下，持有股票的原因
著名的光学专家加盟
股价是否上涨
股价上涨的原因
股价上涨的时间
一个市场对于一个信息集来说称为有效的，
如果不存在利用该信息获得超额利润的机会。
有效和非有效市场中价格对新信息的反应
股票
价格
0 宣布前（ -）或者后（ +）的天数过激反应和回归延迟反应有效市场对新信息的反应
三种有效市场 (efficient capital market)
不同的信息集对证券价格产生影响的速度不一样。
为了处理不同的反应速度，把信息集分成不同的类别。
最常用的一种分类方法：过去价格的信息，
可得的公共信息，所有信息。
针对这三种信息集，有三种形式的有效市场的定义
弱有效市场 (the weak form)
半强有效市场 (the semistrong form)
强有效市场 (strong form)
弱有效市场
定义：一个资本市场称为 弱有效的 或者满足弱有效形式,如果证券价格充分反应了包含在历史价格中的信息。
弱形式有效通常表示成下面的数学形式
+Expected return+Random error
这里
随机误差项的均值为 0，且不同时间的随机误差项是不相关的。
弱形式有效性是最弱类型的有效性。
1 tt PP
股票价格的历史数据是可以免费得到的，
如果这些数据里包含有用的数据，则所有的投资者都会利用它，导致价格调整，
最后，这些数据失去价值。
‘technical’ analysis using past price patterns
will not produce profits.
如果一个市场是弱有效的，考虑一个交易策略
如果某股票的价格连续涨三天，就买进该股票；如果股票的价格连续降三天，就卖出股票。问题：这个策略能否赚钱。
Fairly convinced that markets are weak
form efficient
But,new evidence (i.e.,momentum) has
challenged this.
半强形式有效市场
定义：一个市场是 半强形式有效的,如果价格反应了所有公共可得的信息。
这些信息包括：历史价格数据、与公司生产有关的基本数据、管理的质量、资产负债表、
专利情况、收益预测、会计处理
‘fundamental’ analysis (e.g,sorting through income
statements) will not produce profits.
Ex,Forming portfolios on accounting ratios,balance sheet,
or income statement information will not generate abnormal
profits.
No evidence of abnormal returns after a public
announcement
Professional money managers do not outperform the
market
Market seem to be semi-strong efficient
But,B/M and E/P strategies still challenge this,(may
be risk,may be not)
强形式有效市场
定义：一个市场是 强有效的,如果价格反应了所有的信息，不管是公共的还是私有的。
Insider trading will not produce profits
Ex,Knowing a merger is going to take place
before it is announced publicly.
Although illegal,there is evidence that prices
move before announcements,suggesting
insider trades take place
Insider trading appears profitable,indicating
markets are not strong form efficient.
But,these profits are short lived,suggesting
the market may be close to efficient,
强形式有效性
半强形式有效性
弱形式有效性
说明三种有效性的例子：
总是在股价上涨后卖出股票，能赚到钱
投资者在一家公司宣布增加收益后买该公司股票，能赚到钱
知道采矿公司是否开采到了金子的内部消息后买该公司股票，能赚到钱
4,How can we tell if markets are
inefficient?
Look for stock-picking strategies based
on some past information which have
earned high returns with little risk.
Unfortunately,we can never be sure of
inefficiency.
It is always possible that we are not
measuring risk properly,i.e.,we do not
know what the right discount rate is
This is the ‘Joint hypothesis problem’
5,Why would we expect markets to
be efficient?
the forces of arbitrage
smart investors exploit the mispricing in
securities until it disappears
To show that markets are ineffcient,
need to show
that people make errors in setting prices
that arbitrage fails to eliminate these errors
一些例子
Grossman and Stiglitz的结果
不同市场的有效性不同：中国和美国的股市
小股票和大股票的有效性不同
Random Walks and Efficient Markets
it is often thought that efficient markets prices
move randomly
this is not necessarily true
Strictly speaking,we should characterize stock price as
following a submartingale,meainng that the expected change
in the price can be positive,presumably as compensation for
the time value of the money and systematic risk,Moreover,
the expected return may change over time as risk factors
change.
returns are mean-reverting
if the discount rate for an asset does not change
over time,then it is true that efficient markets
random walk
e.g,over short time frames,returns should look random
6,Evidence for market efficiency
stock prices appear to move randomly
new information appears to be quickly
incorporated into prices
e.g,announcement of a takeover
do an ‘event study’ to look at the stock price
reaction to the news
average over many companies
Professional money managers do not
clearly beat the market on average.
7,Evidence against market efficiency
did the value of the U.S,economy really drop 20% in
October 1987?
the volume of trading on stock exchanges is too high
to be consistent with rational investors
the volatility of the market is too high (Shiller,1982).
why?
why are there so many mutual funds?
there are investment strategies which appear to have
earned higher average returns than is consistent with
their risk
These are so called ‘market anomalies’
suggests that new information is not always immediately
incorporated into prices
This evidence refers to weak/semi-strong versions of market
efficiency
How about strong-form efficiency?
can you make money using inside,non-public information?
YES!
stock prices often move in advance of important company
announcements
from records of insider trades,we find that they make money on
average
BUT:
it's illegal for company insiders (or people
tipped by them) to transact based on
material non-public information
How about institutional investors?
We mentioned money managers do not
exhibit abnormal performance on average,
but may be some local advantage.
8,The market anomalies
Anomalies can be thought of as
investment strategies which seem to
earn high returns without being very
risky
The strategies are normally based on
some firm characteristic:
size of the firm
its price-earnings ratio
Recipe:
form a portfolio based on observable
characteristics,and measure its returns over time
does the strategy give high returns on average?
Why look at average returns?
IF YES:
the strategy may be risky and the high average
returns are just fair compensation for that risk
how do we measure risk? (see below)
if risk does not explain the high returns,is it
evidence of market inefficiency?
it may be spurious,the result of data-mining
if you try many strategies,some of them will do great in
historical data
doesn't tell you anything about future performance
poor risk measurement
frictions to trading and exploiting the anomaly (bid-ask
spread,transactions costs,liquidity,taxes,etc.)
Measuring the risk of a strategy
standard deviation
downside risk
does the strategy sometimes perform very poorly?
beta
looks at strategy's payoff relative to the market's payoff
a high covariance is unattractive (risky)
look at covariance of strategy's payoff with
variables like GNP growth (factor models)
a strategy which does well in good states of the world
and poorly in bad states is risky
Note:
even if none of the above turn up a
measure of risk,efficient markets
enthusiasts will say:
the risk is present,but just hasn't surfaced
within the sample analyzed
OR,we're just not looking at the right measure
of risk
Small vs,Large stocks
Small stocks have outperformed large
stocks by about 12% a year over 1929-
1979 time period
Should we buy lots of small stocks?
Depends on whether they are riskier
small stocks have higher standard deviations
and betas
but not high enough to explain their returns
So,small stocks may be mispriced
BUT:
although small stock returns are high,this
may only be based on a few extraordinary
years
We may be missing some dimension of risk
The January Effect
Much of the small firm effect seems to
occur in January
Much of small firm premium occurs in
first five days of January
often explained by tax-loss selling
however,effect is widespread in international markets also,even
when there's no capital gains tax
and,there still seems to be a size effect after controlling for this.
If the positive January effect is a manifestation of buying pressure,
it should be matched by a symmetric negative December effect.
may not accord with efficient markets
why don't people buy in December in anticipation? and the
predictable January effect flies in the face of efficient market
theory.
‘Window dressing' by institutional investors
infusion of capital at beginning of year
Overreaction studies
some studies suggest that there are
inefficiencies due to people overreacting
to information
（ 1） Losers and Winners---- Long term
reversal
take a three year period and rank stocks on
the basis of their performance over that
period
form a ‘winner" portfolio of the top 10% best-
performing stocks
form a ‘loser" portfolio of the bottom 10% worst-
performing stocks
this is called a contrarian strategy
look at their returns over the next few years
the loser portfolio seems to outperform the
winner portfolio (DeBondt and Thaler (1985))
Two explanations:
overreaction,the winners are firms that people
have become too excited about
subsequently,they realize that they were too optimistic
price falls and returns are low
risk,the losers are riskier firms
their higher returns are just compensation for risk
but losers do not appear riskier on standard measures of
risk
How do we account for risk?
Variance
Use regression analysis and a pricing model
CAPM
FF 3-factor model
Examine when the strategy exhibits the highest
and lowest payoffs
(i.e.,does the strategy do well when the market does?
when we are in the peak of a business cycle? a
recession?,etc.)
Catastrophe risk
（ 2） Value and Growth
form portfolios of value stocks
a value stock is one with low price relative to
some measure of fundamentals,i.e,high
book to market (B/M) ratios
cash flow to price (C/P) ratios
earnings to price (E/P) ratios
also form portfolios of growth stocks,i.e,with
low values of these ratios
Why growth vs,value?
find that value stocks dramatically outperform
growth stocks
Why?
rational,Represents a distress factor in the
economy,Value stocks are more prone to this
source of risk than growth stocks.
higher average returns.
Value stocks are typically `fallen angels'
irrational,Growth stocks are `glamorous',People
tend to want to buy these and stampede towards
them,pushing up the price,and depressing future
returns.
Value stocks have been neglected,causing their price to
fall,and expected returns to rise.
This is an overreaction story.
Also,value stocks do not appear riskier,
however.
they don't have higher variance
they don't have high downside risk (i.e.,do not
underperform often or by that much)
they don't have higher betas
they don't underperform in bad states of the world
Lakonishok,Shleifer,and Vishny (1994)
so maybe it's an inefficiency,driven again by
overreaction
Psychological foundations for
overreaction:
representativeness heuristic
small sample bias
but of course,it could still be risk!
The Fama-French debate
small stocks and high B/M stocks earn
returns that are higher than is required for
their risk,according to the CAPM measure
of risk
two possibilities:
small stocks and high B/M stocks are mispriced
OR,the CAPM isn't measuring risk properly
Fama and French (1993) construct a new
asset-pricing model,the ‘’3-factor model"
which makes small and high B/M stocks
look riskier
start with the market factor,and add two
new factors,F(small)(=SMB) and F(B/M)
(= HML)
they supposedly track good and bad states of
the world
small stocks have high covariance with F(small)
high B/M stocks have high covariance with
F(B/M)
they are riskier!
Evidence in favor of FF:
Adjusting the long-run contrarian profits found by
DeBondt and Thaler (1985) using the 3-factor
model,the profits disappear (i.e.,alpha = 0.).
past long-term losers load higher on SMB and HML than
past winners,even though they have the same market
beta.
long-term losers more risky than long-term winners.
The 3-factor model captures a host of other
anomalies as well!
Evidence against FF:
LSV(1994),HML and other book-to-price ratios
perform well in poor times.
Daniel and Titman (1997):
characteristics rather than factor loadings price assets
better
control for size and B/M characteristics,SMB and HML no
longer explain average returns
ex:
two stocks with same size but different betas on SMB have
the same average return.
two stocks with same beta on SMB but different sizes have
different average returns.
Does this necessarily imply size and BE/ME
are irrational anomalies?
No,could proxy for some true unknown factor,
better captured by the characteristic.
Measurement of betas are prone to estimation
error,which increases noise in the relation
between the betas and average returns.
Underreaction studies
other studies suggest that there are
inefficiencies due to people
underreacting to information
（ 1） Momentum
form portfolios of stocks that performed very well in
recent past (`winners')
i.e,over the past 3 months to one year
similarly,form a portfolio of `loser' stocks
the winners outperform the losers
e.g,buy the winners and sell the losers
e.g,a zero-cost portfolio which buys winners and sells losers
from the past 6 months earns 12% (annually) over the next
6 months!
contrast to the mean-reversion result
source of inefficiency,underreaction to information?
Or,is it risk?
（ 2） Earnings Announcements
slow price response to earnings
announcements (post-earnings
announcement drift)
each quarter,rank stocks on the size of
the surprise in their earnings
announcement (surprise = actual -
expected)
form a portfolio of stocks with largest
positive surprises (portfolio A)
and a portfolio of stocks with the
largest negative surprises (portfolio B)
A outperforms B
is it risk?
beta,factor model type checks don't find
any
over 50 quarters from 1974-1986,strategy
earned positive abnormal returns 46 times
Many examples of underreaction to
information:
to changes in dividend policy
strategy,buy companies that have just announced a
dividend increase; sell those that have just announced a
cut in dividends
to repurchases of shares
strategy,buy companies that have just announced a
share repurchase
Psychological Foundations for Underreaction:
conservatism,pessimism
Other anomalies
The new issues puzzle
Hot offering
If these anomalies are inefficiencies,
why doesn't arbitrage eliminate them?
in practice,arbitrage is limited
do not have infinite number of stocks,therefore
there is some risk
sometimes the strategy can do very poorly,and
you lose a lot of money
money managers and individuals may have short
horizons (due to regular evaluations or
psychological preferences)
a mispricing can take a while to close
in fact it may not close within the investor's horizon
investor will restrict the size of position taken
there can be high transactions and trading costs
due to turnover in these strategies
liquidity can be low at times when you need it most
monitoring can be high
there are short sales constraints
Measurement issues,These anomalies are
market-based measures
e.g.,they include price
size,BE/ME,past returns (contrarian and momentum),
E/P,etc.,all contain price in them.
so anything missed by the pricing
model will show up in one of these variables
e.g.,appears on both sides of the regression
equation
this was noted by Ball (1978) and formalized in Berk
(1995)
we may never know if picking up mispricing or
inadequacy of pricing model.
111t tt P DPtr
tP
Data mining
techniques of finding anomalies are
often subject to the data mining critique
if you try enough variables,something will
eventually appear to predict returns
but,the forecast power of this variable will
be completely spurious (i.e.,it won't work
out of sample)
e.g.,generate 100 different data series of
completely random numbers
run a regression of actual stock returns on each
of the 100 random data series.
some of the regressions will produce significant
results:
does this mean you can make money? NO!
the data mining critique is very powerful:
bear it in mind when people try to impress you with
strategies that worked great in the past.
they probably won't work in the future!
Response to Data Mining Critique:
Many of these anomalies appear in other
(international) markets (see Hawawini and
Keim (1995) article)
Acid Test:
Do the strategies designed to exploit the
anomalies perform well in the future?
Only time will tell.
So,is the market efficient?
Behavior finance
The traditional finance paradigm seeks to understand
financial markets using models in which agents are
‘rational’
Bayes law,when they receive new information,agent
update their beliefs correctly
Expected utility function,given their beliefs,agents make
choices that are normatively acceptable
Behavior finance analyzes what happens when we
relax one or both,of the two tenets that underlie
individual rationality,The two buildings blocks of
behavior finance:
Limits to arbitrage
psychology.
9,市场是有效的吗？
Are markets efficient? We ought instead
to ask a more quantitative question,
How efficient are market?
实际的市场具有半强形式有效性。
关于市场有效性的一些错误观点
the efficacy of dart throwing
all the efficient-market hypothesis really
says is that,on average,the manager
will not be able to achieve an abnormal
or excess return.
The efficient market protects the sheep
from the wolves,but nothing can protect
the sheep from themselves.
What efficiency does say is that the price that a
firm will obtain when it sells a share of its stock
is a fair price in the sense that it reflects the
value of that stock given the information that is
available about it,Shareholders need not worry
that they are paying too much for a stock with
a low dividend or some other characteristic,
because the market has already incorporated it
into the price,However,investors still have to
worry about such things as their lever of risk
exposure and their degree of diversification.
Price fluctuations
A stock in an efficient market adjusts to new
information by changing price,The absence of
price movements in a changing world might
suggest an inefficiency.
Stockholder Disinterest
有效市场中证券组合管理的功能
由于年龄、税基、风险回避、就业等不同，
投资者选择最适合自己的投资组合。
有效市场中证券组合管理的功能是满足具体需要，而不是 beating the market.
效率市场的定义
效率市场的实证结果
证券市场中的几种反常现象
行为金融
Maurice Kendall,The analysis of
economic time series,Part I,Prices,
Journal of the Royal statistical society
96,1953.
He could identify no predictable patterns
in stock prices.
1,Why are not stock prices
predictable?
例子：一种模型预测股票价格三天后将从 20元 /股涨至 23元 /股
The forecast of a future price increase
will lead instead to an immediate price
increase
the stock price will immediately reflect
the good news implicit in the model’s
forecast.
例子,2001年 10月 21日晚，证监会宣布暂停国有股减持
例子,2002年 4月 5日宣布，从 5月 1日起交易佣金使用 0.3%上限向下浮动制。
例子：中国人民银行决定，从 2004年 10月 29日起上调金融机构存贷款基准利率并放宽人民币贷款利率浮动区间和允许人民币存款利率下浮。
金融机构一年期存款基准利率上调 0.27个百分点，由现行的 1.98%提高到 2.25%，一年期贷款基准利率上调 0.27个百分点，由现行的
5.31%提高到 5.58%。
由于存在聪明的投资者的原因，任何能够用来对股票价格作预测的信息已经反映在股票的价格中。
任何新信息，如果是可以预测的，则已经反映在价格中；如果是不可预测的，
则导致的股票价格变动也是不可预测的，
即随机的信息导致随机的股票价格变化。
2,股票价格的随机游走 (random walk)
股票价格的变动服从随机游走的形式。
随机游走的形式并不是说明市场是非理性的，而恰恰表明这是投资者真相寻求相关信息，以使得自己在别的投资者获得这种信息之前买或者卖股票而获得利润的结果。
反之，如果股票价格不是随机变动而是可以预测的，则表明所有的相关信息并没有完全反映在价格中，这表明市场是非有效的。
股票价格反映了所有可得信息称为 有效性市场假设 。
有效资本市场指的是现时市场价格能够反映可得信息的资本市场，在这个市场中，不存在利用可得信息获得超额利润的机会。
资本市场的有效性这一概念对于金融投资者而言具有非常重要的意义。因为市场的有效性消除了许多可以提高收益的策略。
Market efficiency means
Prices are correct
They fully reflect all available information
People use all available information in forming expectations
about future cash flows.
The discount rate is right for the riskness of the cash flows.
Prices react to new information quickly and to the right extent
There is no free lunch
The only way you can get higher returns is by taking on more risk
There is no information out there that can be used to construct
strategies that earn returns higher than required for their risk.
When we say ‘prices are correct’,we are implicitly statement what
‘correct’ is (i.e.,we are assuming an asset pricing model)
1 1i i
ittt
r
dEP
3,有效资本市场的描述
例子,F-stop Camera Corporation is
attempting to develop a camera that
will double the speed of the auto--
focusing system now available,
在特定的价格下，持有股票的原因
著名的光学专家加盟
股价是否上涨
股价上涨的原因
股价上涨的时间
一个市场对于一个信息集来说称为有效的，
如果不存在利用该信息获得超额利润的机会。
有效和非有效市场中价格对新信息的反应
股票
价格
0 宣布前（ -）或者后（ +）的天数过激反应和回归延迟反应有效市场对新信息的反应
三种有效市场 (efficient capital market)
不同的信息集对证券价格产生影响的速度不一样。
为了处理不同的反应速度，把信息集分成不同的类别。
最常用的一种分类方法：过去价格的信息，
可得的公共信息，所有信息。
针对这三种信息集，有三种形式的有效市场的定义
弱有效市场 (the weak form)
半强有效市场 (the semistrong form)
强有效市场 (strong form)
弱有效市场
定义：一个资本市场称为 弱有效的 或者满足弱有效形式,如果证券价格充分反应了包含在历史价格中的信息。
弱形式有效通常表示成下面的数学形式
+Expected return+Random error
这里
随机误差项的均值为 0，且不同时间的随机误差项是不相关的。
弱形式有效性是最弱类型的有效性。
1 tt PP
股票价格的历史数据是可以免费得到的，
如果这些数据里包含有用的数据，则所有的投资者都会利用它，导致价格调整，
最后，这些数据失去价值。
‘technical’ analysis using past price patterns
will not produce profits.
如果一个市场是弱有效的，考虑一个交易策略
如果某股票的价格连续涨三天，就买进该股票；如果股票的价格连续降三天，就卖出股票。问题：这个策略能否赚钱。
Fairly convinced that markets are weak
form efficient
But,new evidence (i.e.,momentum) has
challenged this.
半强形式有效市场
定义：一个市场是 半强形式有效的,如果价格反应了所有公共可得的信息。
这些信息包括：历史价格数据、与公司生产有关的基本数据、管理的质量、资产负债表、
专利情况、收益预测、会计处理
‘fundamental’ analysis (e.g,sorting through income
statements) will not produce profits.
Ex,Forming portfolios on accounting ratios,balance sheet,
or income statement information will not generate abnormal
profits.
No evidence of abnormal returns after a public
announcement
Professional money managers do not outperform the
market
Market seem to be semi-strong efficient
But,B/M and E/P strategies still challenge this,(may
be risk,may be not)
强形式有效市场
定义：一个市场是 强有效的,如果价格反应了所有的信息，不管是公共的还是私有的。
Insider trading will not produce profits
Ex,Knowing a merger is going to take place
before it is announced publicly.
Although illegal,there is evidence that prices
move before announcements,suggesting
insider trades take place
Insider trading appears profitable,indicating
markets are not strong form efficient.
But,these profits are short lived,suggesting
the market may be close to efficient,
强形式有效性
半强形式有效性
弱形式有效性
说明三种有效性的例子：
总是在股价上涨后卖出股票，能赚到钱
投资者在一家公司宣布增加收益后买该公司股票，能赚到钱
知道采矿公司是否开采到了金子的内部消息后买该公司股票，能赚到钱
4,How can we tell if markets are
inefficient?
Look for stock-picking strategies based
on some past information which have
earned high returns with little risk.
Unfortunately,we can never be sure of
inefficiency.
It is always possible that we are not
measuring risk properly,i.e.,we do not
know what the right discount rate is
This is the ‘Joint hypothesis problem’
5,Why would we expect markets to
be efficient?
the forces of arbitrage
smart investors exploit the mispricing in
securities until it disappears
To show that markets are ineffcient,
need to show
that people make errors in setting prices
that arbitrage fails to eliminate these errors
一些例子
Grossman and Stiglitz的结果
不同市场的有效性不同：中国和美国的股市
小股票和大股票的有效性不同
Random Walks and Efficient Markets
it is often thought that efficient markets prices
move randomly
this is not necessarily true
Strictly speaking,we should characterize stock price as
following a submartingale,meainng that the expected change
in the price can be positive,presumably as compensation for
the time value of the money and systematic risk,Moreover,
the expected return may change over time as risk factors
change.
returns are mean-reverting
if the discount rate for an asset does not change
over time,then it is true that efficient markets
random walk
e.g,over short time frames,returns should look random
6,Evidence for market efficiency
stock prices appear to move randomly
new information appears to be quickly
incorporated into prices
e.g,announcement of a takeover
do an ‘event study’ to look at the stock price
reaction to the news
average over many companies
Professional money managers do not
clearly beat the market on average.
7,Evidence against market efficiency
did the value of the U.S,economy really drop 20% in
October 1987?
the volume of trading on stock exchanges is too high
to be consistent with rational investors
the volatility of the market is too high (Shiller,1982).
why?
why are there so many mutual funds?
there are investment strategies which appear to have
earned higher average returns than is consistent with
their risk
These are so called ‘market anomalies’
suggests that new information is not always immediately
incorporated into prices
This evidence refers to weak/semi-strong versions of market
efficiency
How about strong-form efficiency?
can you make money using inside,non-public information?
YES!
stock prices often move in advance of important company
announcements
from records of insider trades,we find that they make money on
average
BUT:
it's illegal for company insiders (or people
tipped by them) to transact based on
material non-public information
How about institutional investors?
We mentioned money managers do not
exhibit abnormal performance on average,
but may be some local advantage.
8,The market anomalies
Anomalies can be thought of as
investment strategies which seem to
earn high returns without being very
risky
The strategies are normally based on
some firm characteristic:
size of the firm
its price-earnings ratio
Recipe:
form a portfolio based on observable
characteristics,and measure its returns over time
does the strategy give high returns on average?
Why look at average returns?
IF YES:
the strategy may be risky and the high average
returns are just fair compensation for that risk
how do we measure risk? (see below)
if risk does not explain the high returns,is it
evidence of market inefficiency?
it may be spurious,the result of data-mining
if you try many strategies,some of them will do great in
historical data
doesn't tell you anything about future performance
poor risk measurement
frictions to trading and exploiting the anomaly (bid-ask
spread,transactions costs,liquidity,taxes,etc.)
Measuring the risk of a strategy
standard deviation
downside risk
does the strategy sometimes perform very poorly?
beta
looks at strategy's payoff relative to the market's payoff
a high covariance is unattractive (risky)
look at covariance of strategy's payoff with
variables like GNP growth (factor models)
a strategy which does well in good states of the world
and poorly in bad states is risky
Note:
even if none of the above turn up a
measure of risk,efficient markets
enthusiasts will say:
the risk is present,but just hasn't surfaced
within the sample analyzed
OR,we're just not looking at the right measure
of risk
Small vs,Large stocks
Small stocks have outperformed large
stocks by about 12% a year over 1929-
1979 time period
Should we buy lots of small stocks?
Depends on whether they are riskier
small stocks have higher standard deviations
and betas
but not high enough to explain their returns
So,small stocks may be mispriced
BUT:
although small stock returns are high,this
may only be based on a few extraordinary
years
We may be missing some dimension of risk
The January Effect
Much of the small firm effect seems to
occur in January
Much of small firm premium occurs in
first five days of January
often explained by tax-loss selling
however,effect is widespread in international markets also,even
when there's no capital gains tax
and,there still seems to be a size effect after controlling for this.
If the positive January effect is a manifestation of buying pressure,
it should be matched by a symmetric negative December effect.
may not accord with efficient markets
why don't people buy in December in anticipation? and the
predictable January effect flies in the face of efficient market
theory.
‘Window dressing' by institutional investors
infusion of capital at beginning of year
Overreaction studies
some studies suggest that there are
inefficiencies due to people overreacting
to information
（ 1） Losers and Winners---- Long term
reversal
take a three year period and rank stocks on
the basis of their performance over that
period
form a ‘winner" portfolio of the top 10% best-
performing stocks
form a ‘loser" portfolio of the bottom 10% worst-
performing stocks
this is called a contrarian strategy
look at their returns over the next few years
the loser portfolio seems to outperform the
winner portfolio (DeBondt and Thaler (1985))
Two explanations:
overreaction,the winners are firms that people
have become too excited about
subsequently,they realize that they were too optimistic
price falls and returns are low
risk,the losers are riskier firms
their higher returns are just compensation for risk
but losers do not appear riskier on standard measures of
risk
How do we account for risk?
Variance
Use regression analysis and a pricing model
CAPM
FF 3-factor model
Examine when the strategy exhibits the highest
and lowest payoffs
(i.e.,does the strategy do well when the market does?
when we are in the peak of a business cycle? a
recession?,etc.)
Catastrophe risk
（ 2） Value and Growth
form portfolios of value stocks
a value stock is one with low price relative to
some measure of fundamentals,i.e,high
book to market (B/M) ratios
cash flow to price (C/P) ratios
earnings to price (E/P) ratios
also form portfolios of growth stocks,i.e,with
low values of these ratios
Why growth vs,value?
find that value stocks dramatically outperform
growth stocks
Why?
rational,Represents a distress factor in the
economy,Value stocks are more prone to this
source of risk than growth stocks.
higher average returns.
Value stocks are typically `fallen angels'
irrational,Growth stocks are `glamorous',People
tend to want to buy these and stampede towards
them,pushing up the price,and depressing future
returns.
Value stocks have been neglected,causing their price to
fall,and expected returns to rise.
This is an overreaction story.
Also,value stocks do not appear riskier,
however.
they don't have higher variance
they don't have high downside risk (i.e.,do not
underperform often or by that much)
they don't have higher betas
they don't underperform in bad states of the world
Lakonishok,Shleifer,and Vishny (1994)
so maybe it's an inefficiency,driven again by
overreaction
Psychological foundations for
overreaction:
representativeness heuristic
small sample bias
but of course,it could still be risk!
The Fama-French debate
small stocks and high B/M stocks earn
returns that are higher than is required for
their risk,according to the CAPM measure
of risk
two possibilities:
small stocks and high B/M stocks are mispriced
OR,the CAPM isn't measuring risk properly
Fama and French (1993) construct a new
asset-pricing model,the ‘’3-factor model"
which makes small and high B/M stocks
look riskier
start with the market factor,and add two
new factors,F(small)(=SMB) and F(B/M)
(= HML)
they supposedly track good and bad states of
the world
small stocks have high covariance with F(small)
high B/M stocks have high covariance with
F(B/M)
they are riskier!
Evidence in favor of FF:
Adjusting the long-run contrarian profits found by
DeBondt and Thaler (1985) using the 3-factor
model,the profits disappear (i.e.,alpha = 0.).
past long-term losers load higher on SMB and HML than
past winners,even though they have the same market
beta.
long-term losers more risky than long-term winners.
The 3-factor model captures a host of other
anomalies as well!
Evidence against FF:
LSV(1994),HML and other book-to-price ratios
perform well in poor times.
Daniel and Titman (1997):
characteristics rather than factor loadings price assets
better
control for size and B/M characteristics,SMB and HML no
longer explain average returns
ex:
two stocks with same size but different betas on SMB have
the same average return.
two stocks with same beta on SMB but different sizes have
different average returns.
Does this necessarily imply size and BE/ME
are irrational anomalies?
No,could proxy for some true unknown factor,
better captured by the characteristic.
Measurement of betas are prone to estimation
error,which increases noise in the relation
between the betas and average returns.
Underreaction studies
other studies suggest that there are
inefficiencies due to people
underreacting to information
（ 1） Momentum
form portfolios of stocks that performed very well in
recent past (`winners')
i.e,over the past 3 months to one year
similarly,form a portfolio of `loser' stocks
the winners outperform the losers
e.g,buy the winners and sell the losers
e.g,a zero-cost portfolio which buys winners and sells losers
from the past 6 months earns 12% (annually) over the next
6 months!
contrast to the mean-reversion result
source of inefficiency,underreaction to information?
Or,is it risk?
（ 2） Earnings Announcements
slow price response to earnings
announcements (post-earnings
announcement drift)
each quarter,rank stocks on the size of
the surprise in their earnings
announcement (surprise = actual -
expected)
form a portfolio of stocks with largest
positive surprises (portfolio A)
and a portfolio of stocks with the
largest negative surprises (portfolio B)
A outperforms B
is it risk?
beta,factor model type checks don't find
any
over 50 quarters from 1974-1986,strategy
earned positive abnormal returns 46 times
Many examples of underreaction to
information:
to changes in dividend policy
strategy,buy companies that have just announced a
dividend increase; sell those that have just announced a
cut in dividends
to repurchases of shares
strategy,buy companies that have just announced a
share repurchase
Psychological Foundations for Underreaction:
conservatism,pessimism
Other anomalies
The new issues puzzle
Hot offering
If these anomalies are inefficiencies,
why doesn't arbitrage eliminate them?
in practice,arbitrage is limited
do not have infinite number of stocks,therefore
there is some risk
sometimes the strategy can do very poorly,and
you lose a lot of money
money managers and individuals may have short
horizons (due to regular evaluations or
psychological preferences)
a mispricing can take a while to close
in fact it may not close within the investor's horizon
investor will restrict the size of position taken
there can be high transactions and trading costs
due to turnover in these strategies
liquidity can be low at times when you need it most
monitoring can be high
there are short sales constraints
Measurement issues,These anomalies are
market-based measures
e.g.,they include price
size,BE/ME,past returns (contrarian and momentum),
E/P,etc.,all contain price in them.
so anything missed by the pricing
model will show up in one of these variables
e.g.,appears on both sides of the regression
equation
this was noted by Ball (1978) and formalized in Berk
(1995)
we may never know if picking up mispricing or
inadequacy of pricing model.
111t tt P DPtr
tP
Data mining
techniques of finding anomalies are
often subject to the data mining critique
if you try enough variables,something will
eventually appear to predict returns
but,the forecast power of this variable will
be completely spurious (i.e.,it won't work
out of sample)
e.g.,generate 100 different data series of
completely random numbers
run a regression of actual stock returns on each
of the 100 random data series.
some of the regressions will produce significant
results:
does this mean you can make money? NO!
the data mining critique is very powerful:
bear it in mind when people try to impress you with
strategies that worked great in the past.
they probably won't work in the future!
Response to Data Mining Critique:
Many of these anomalies appear in other
(international) markets (see Hawawini and
Keim (1995) article)
Acid Test:
Do the strategies designed to exploit the
anomalies perform well in the future?
Only time will tell.
So,is the market efficient?
Behavior finance
The traditional finance paradigm seeks to understand
financial markets using models in which agents are
‘rational’
Bayes law,when they receive new information,agent
update their beliefs correctly
Expected utility function,given their beliefs,agents make
choices that are normatively acceptable
Behavior finance analyzes what happens when we
relax one or both,of the two tenets that underlie
individual rationality,The two buildings blocks of
behavior finance:
Limits to arbitrage
psychology.
9,市场是有效的吗？
Are markets efficient? We ought instead
to ask a more quantitative question,
How efficient are market?
实际的市场具有半强形式有效性。
关于市场有效性的一些错误观点
the efficacy of dart throwing
all the efficient-market hypothesis really
says is that,on average,the manager
will not be able to achieve an abnormal
or excess return.
The efficient market protects the sheep
from the wolves,but nothing can protect
the sheep from themselves.
What efficiency does say is that the price that a
firm will obtain when it sells a share of its stock
is a fair price in the sense that it reflects the
value of that stock given the information that is
available about it,Shareholders need not worry
that they are paying too much for a stock with
a low dividend or some other characteristic,
because the market has already incorporated it
into the price,However,investors still have to
worry about such things as their lever of risk
exposure and their degree of diversification.
Price fluctuations
A stock in an efficient market adjusts to new
information by changing price,The absence of
price movements in a changing world might
suggest an inefficiency.
Stockholder Disinterest
有效市场中证券组合管理的功能
由于年龄、税基、风险回避、就业等不同，
投资者选择最适合自己的投资组合。
有效市场中证券组合管理的功能是满足具体需要，而不是 beating the market.
