Arbitrage pricing theory (APT): an asset pricing theory that is derived from a factor model,

using diversification and arbitrage arguments. The theory describes the relationship between
expected returns on securities, given that there are no opportunities to create wealth through risk
free arbitrage opportunities.
The APT predicts a security market line linking expected returns to risk, through a different path.
The theory relies on three main propositions:
Security returns can be described by a factor model;
There are sufficient securities to diversify away idiosyncratic risk;
Well-functioning security markets do not allow for the persistence of arbitrage

Law of one price: the rule stipulating that securities or portfolios with equal returns under all
circumstances must sell at equal prices to preclude arbitrage opportunities. Revenues and
expenses are listed and their difference is calculated as net income.
The critical property of a risk-free arbitrage portfolio is that any investor, regardless of risk
aversion or wealth, will want to take an infinite position in it.
There is an important difference between arbitrage and risk-return dominance arguments in
support of equilibrium:
A dominance argument holds that when an equilibrium price relationship is violated, many
investors will make limited portfolio changes, depending on their degree of risk aversion.
Aggregation of these limited portfolio changes is required to create a large volume of buying and
selling, which in turn restores equilibrium prices.
By contrast, when arbitrage opportunities exist each investor wants to take as large a position as
possible; hence it will not take many investors to bring about the price pressures necessary to
restore equilibrium. Therefore, implications for prices derived from no-arbitrage arguments are
stronger than implications derived from a risk- return dominance argument.

Risk arbitrage : speculation on perceived mispriced securities, usually in connection with

merger and acquisitions targets.
If beta isnt the only source of risk that is priced, then CAPM-alphas should persist. A stock
might be cheap for a good reason.
 When you compare two assets, even if they arent identical, their prices should still be highly
correlated; i.e. a stock that is highly correlated to small stocks should generate positive alpha.
In other words, any stock that shares the same risks as small-caps should have similar pricing.
The set of portfolios for which the nonsystematic variance approaches zero as n gets large
consists of more portfolios that just the equally weighted portfolio.
Well-diversified portfolio a portfolio spread out over may securities in such a way that the
weight in any one security is close to zero.

Imagine a single factor market where the well-diversified portfolio, M, represents the market
factor, F, the excess return on any security is given by:
= + +
and that of a well-diversified (therefore zero residual) portfolio, P, is:
= +
Although the APT is built on the foundation of well-diversified portfolios, weve seen that
even large portfolios may have non-negligible residual risk.

Since arbitrage activity will quickly pin the risk premium of any zero-beta well-diversified
portfolio to zero, then that for any well-diversified P, =

In other words, the risk premium (expected excess return) on portfolio P is the product of its
beta and the market index risk premium. The previous equation this establish that SML of the
CAPM applies to well-diversified portfolios simply by virtue of the "no-arbitrage"
requirement of the APT.

If residual risk is sufficiently high and the obstacles to complete diversification are considerably
difficult, we cannot have full confidence in the APT and the arbitrage activities that rely on it.
The previous equation might also be used to predict the risk premiums of portfolios with residual
risk; the higher the residual risk is the less accurate is the approximation.
The APT is more efficient than the CAPM in terms that it does not require that all investors must
be mean-variance optimizers. It is sufficient that a small number of sophisticated arbitrageurs
 scour the market for arbitrage opportunities. This alone produces a mean return-beta (previous
equation) that is a good and unbiased approximation for all assets but those with significant
residual risk.
The reason APT is not fully superior to the CAPM is that at the level of individual assets and
high residual risk, pure arbitrage may be insufficient to enforce the expected risk premium
equation. Therefore, we need to run to the CAPM as the theoretical construct behind
equilibrium risk premiums.
Comparing the APT arbitrage strategy to maximization of the Sharpe ratio in the contex of an
index modl may well be the more useful framework for analysis.
The APT is couched in a single-factor market, and applies with perfect accuracy to well-
diversified portfolios. It shows arbitrageurs how to generate infinite profits if the risk premium
of a well-diversified portfolio deviates from the risk premium equation. The trades executed
by these arbitrageurs are the enforcers of the accuracy of this equation. The APT ignores the
fact that and increase in a position of a no fully risk free (well-diversified) will increase the
risk of the 'arbitrage' position, potentially without bound. But if you limit the scale of the risky
arbitrage, the composition of your overall risky portfolio then becomes relevant.
There are multiple factors that may influence the expected returns on stock, such as the interest
rate fluctuation, inflation rates and so on. Exposure to any of these factors will affect a stock's
risk and hence its expected return. We can derive a multifactor version of the APT to
accommodate these multiple sources of risk, a two-factor model would be:
= + 1 1 + 2 2 +
Each factor has zero expected value because each measures the surprise in the systematic
variable rather than the level of the variable. Similarly, the firm-specific component of
unexpected return also has zero expected value. Factor portfolio- a well diversified portfolio
constructed to have a beta of 1.0 on one factor and a beta of zero on any other factor. The
returns on the factor portfolio track the evolution of particular sources of macroeconomic risk,
but are uncorrelated with other sources of risk.

Factor portfolios will serve as the bench mark portfolios for a multifactor security market line.
The factor exposures of any portfolio, P, are given by its betas 1 2. A competing
portfolio, Q, can be formed by investing in factor portfolios with the following weight: 1 in
 the first factor portfolio, 2 in the second factor portfolio, and 1- 1 - 2 in T-bills. By
construction, portfolio, Q will have betas equal to those of portfolio P and expected return of:
= + 1 (1 ) + 2 (2 )

When choosing factors we need to follow these two guidelines:

First, we want to limit ourselves to systematic factors with considerable ability to explain
security returns.
Second, we wish to choose factors that seem likely to be important risk factors, that is , factors
that concern investors sufficiently that they will demand meaningful risk premium to bear
exposure to those sources of risk.
One example of the multifactor approach is the work of Chen, Roll and Ross, who chose the
following set of factors on the basis of the ability of these factors to paint a broad picture of
the macro economy.
IP = % change in the industrial production
EI= % change in the expected inflation
UI= % change in the unanticipated inflation
CG= Excess return of long term corporate bonds over long term government bonds
GB= Excess return of long term government bonds over T-bills
This list gives rise to the following five-factor model of security returns during holding period as
a function of the change in the set of macroeconomic indicators:

To estimate the betas of a given stock we would use a multiple regression of the returns of the
stock in each period on the five macroeconomic factors. The residual variance of the
regression estimates the firm-specific risk.
An alternative approach to specifying macroeconomic factors as candidates for relevant sources
of systematic risk uses firm characteristics that seem on empirical ground to proxy for
exposure to systematic risk. In other words, the factors are chosen as variable that on past
evidence seem to predict high average returns and therefore may be capturing risk premiums.
One example of this approach is the so-called Fama and French (FF) three- factor model.