Arbitrage Pricing Theory

Arbitrage Pricing Theory (APT) was proposed by Stephen S.Rose and presented in his article „The
arbitrage theory of Capital Asset Pricing “, published in Journal of Economic Theory in 1976. In the
CAPM returns on individual assets are related to returns on the market. The APT states that the
market return is decided by several varied factors that affect different securities differently. These
factors can be fundamental factors or statistical.

In APT, asset valuation and its returns can be predicted by a linear relationship between asset's
expected return and some macro-economic factors that affect the asset's risk. Arbitrage pricing
theory offers analysts and investors a multi-factor pricing model for securities based on the
relationship between a financial asset's expected return and its risks.

The theory aims to predict the fair market price of a security that may be temporarily mispriced.
The theory assumes that market action is not always perfectly efficient, and therefore it occasionally
results in assets being mispriced either overvalued or undervalued for a brief period. However,
market action should eventually correct the situation, moving price back to its fair market value. To
an arbitrageur, temporarily mispriced securities are a short-term opportunity to profit risk-free,
hence the name Arbitrage Pricing Theory.

The APT is a more flexible and complex alternative to CAPM. The theory supplies investors and
analysts' methods to customize their research. However, it is more difficult to apply, as it is a multi-
factor model. It takes considerable amount of time and effort to decide all risk factors that may
influence the price of an asset.

The APT is based upon the assumption that there are a few major macro-economic factors that
influence security returns.

APT holds that

(i) the expected returns from a security are a linear function of several factors affecting the returns
from the securities in the market; these factors may be interest rate, inflation, currency rates, GDP
growth, oil prices etc., each factor is represented by a factor specific beta coefficient.

(ii) the correct price of the asset can be decided using the APT model derived expected rate.

This theory is an alternative to CAPM. CAPM assumes that the expected returns from security is
related to only one specific factor, i.e., market returns. But market itself is affected by various
macro factors. The CAPM considers the effect of these factors in totality i.e., in the form of market
returns. The APT considers these factors separately. These macro-economic factors are also referred
to as risk factors.

Page No 1 of 8
 Arbitrage Pricing Theory
The APT is a multiple factor model. Students should not confuse it with multiple index model i.e.,
Markowitz model. Under Markowitz model each security in the portfolio was compared with all
other Securities in the portfolio. Whereas, under APT a stock is not compared with other stocks in
the portfolio, instead multiple risk factors affecting the stock are decided and the impact of changes
in such factors on stock is evaluated and denoted in form of factor specific beta. Thus, Markowitz
model is a multiple index model while APT is multiple factor model and they both are not same.

APT in its original form had only four risk factors:

(i) Change in Inflation,

(ii) Change in levels of industrial production,

(iii) Shifts in risk Premium and

(iv) change in the shape of the term structure of interest rates.

But analysts can identify more risk factors which affect the stock prices and analyze their impact on
returns of stock.

Features of APT

1. Multi Factor Model

As per APT, a security's returns are not affected by a single macro-economic variable (also called as
risk factors), but different risk factors affect company in a diverse ways, therefore effect of each
variable on company should be evaluated separately. Hence APT is a multi-factor model (and not a
multiple index model)

2. Alternative to CAPM

CAPM compares security behavior with a general economic factor i.e., the stock market. As against
this, APT studies behavior of a security with multiple and more specific macro-economic variable.
Hence, APT helps an analyst evaluate and study systematic risk better than CAPM.

3. Linear Model

APT is a linear model, i.e., expected returns of a security have linear relationship with various risk
factors and therefore can be analyzed with help of statistical techniques such as regression analysis
and characteristics line of each such factor.

4. Does Not Identify Risk Factors

APT helps analyze the impact of a risk factor on a stock's return, but it does not help an analyst to
identify risk factors. For example, an analyst wants to evaluate returns of say ONGC Ltd. APT will
help the analyst understand the impact of change in oil prices, government import policy,
Page No 2 of 8
 Arbitrage Pricing Theory
fluctuation in foreign exchange, etc. But APT will not help identify the risk factors itself. This
means that the analyst must identify risk factors such as oil prices, fluctuation in foreign exchange,
etc. himself and then use APT to analyze their impact on stock price individually.

5. Based on Law of One Price

Since two identical assets cannot sell at different prices, equilibrium prices adjust to eliminate all
arbitrage opportunities. In other words, arbitrage opportunity does not exist, and even if it exists, it
will disappear quickly.

6. Unsystematic Risk can be Diversified Away

AS in case of CAPM, APT assumes that unsystematic or firm specific can be diversified away.

7. Non-Dependent on Market Portfolio

APT evaluates all macro variables individually, it does not depend on market portfolio.

Assumptions Under APT

1. Diversification works

Unsystematic risk or firm specific risk can be diversified away.

2. Returns are Generated Using Factor Model

APT assumes that asset returns are dependent on systematic factors where each such factor affects
the company differently.

3. No Arbitrage Exists

No arbitrage opportunities exist among well-diversified portfolios. investors. If any arbitrage

4. Profit Intent

All market participants trade with the intent of profit maximization.

5. Frictionless Markets

There is no or very low transaction cost, no taxes, short selling is possible and an infinite number of
securities is available.

Page No 3 of 8
 Arbitrage Pricing Theory
Advantages of APT

1. Better Accuracy

Unlike CAPM which assumes that market reflects all macro-economic variables, APT evaluates
impact of each variable individually. This makes on APT a more detailed study of factors affecting a
stock price, hence providing more right price as compared to CAPM.

2. Lesser Assumptions

The APT does not make same assumptions about individual portfolios as other predictive theories.
It also has fewer restrictions on the information used for predictions. Because there is more
information available, with fewer restrictions, the results are more reliable than with other
competitive models.

3. More Flexible about Risk Factors

Although APT does not use specific factors like other pricing models, there are four crucial factors
that are considered by the theory. APT looks at changes in inflation, changes in industrial
production, shifts in risk premiums, and shifts in the structure of interest rates, when creating long-
term predictive models. Apart from this, an analyst can identify other macro-economic factors that
affect the company.

4. It Allows for Unanticipated Changes

CAPM is based on the idea that no surprises are going to happen. That is an unrealistic expectation,
so Ross included a variable in the equation to account for any unanticipated change. That makes it
easier for investors to identify assets which have the strongest potential for growth or the strongest
potential for failure, based on the information that is provided by any unanticipated opportunity.

5. Helps Investors Find Arbitrage Opportunities

The goal of APT is to help investors find securities in the market that mispriced. Once these are
identified, it is possible to build a portfolio based on them to generate higher than normal returns. If
an undervalued portfolio is identified, such opportunities can be exploited to generate profits.

Limitations of the APT

1. Too Much Data to Analyse

As each risk factor is evaluated individually, the data analysis can be quite overwhelming. Enough
risk factors should be used in the APT equation for forecast to be right. Using too few risk factors is
not a good plan. At the same time, many risk factors and a lot of data analysis does not imply better
APT model.

Page No 4 of 8
 Arbitrage Pricing Theory
2. It Requires Risk Factors to be Accurate

The APT does not identify risk for the analyst, it only evaluates it. For APT to be useful, it requires
investors to have a clear perception of the risk, as well as the source of that risk. Only then, will this
theory be able to factor in reasonable estimates with good factoring sensitivities for a higher level of
accuracy in prediction. If there is no clear definition of a risk and its source, the effectiveness of the
predictive qualities that APT will diminish.

3. It Requires the Portfolio to be Examined as a Single Unit

The APT is only useful when examining a single item of investment. Because of this examining an
entire portfolio with diverse investments is difficult under the APT. Therefore, the entire portfolio is
examined as a single investment the APT. This approach does not evaluate every security separately
and some assumptions are made for the evaluation. That can reduce the accuracy of the outcomes.

4. Misleading Name

Arbitrage means riskless profit. Although APT uses the word arbitrage, it does not guarantee that
profits will happen. There are securities which are undervalued at current date for reasons that fall
outside of the scope of the APT. Also, some risk factors are not "real" risks, as they are built into the
pricing mechanisms by the investors, who have a certain fear of specific market conditions.

5. High Maintenance

The factors that affect the stock price for a particular stock may change over a period. Moreover, the
sensitivities associated may also undergo shifts which need to be continuously checked making it
difficult to calculate and maintain.

RETURNS UNDER APT

APT evaluates each macro-economic factor or risk factor separately.

This involves following steps:

1. Risk Free Rate

It is denoted by Greek alphabet Lambda, ƛ0 , i.e., Lambda = Zero

2. Risk Premium

Risk premium for each risk factor is also denoted by Greek Alphabet Lambda (ƛ)

ƛ1 for risk factor 1, ƛ2 for risk factor 2 and so on.

Risk premium is an independent factor and does not depend on the investment being valued.

Page No 5 of 8
 Arbitrage Pricing Theory
3. Beta (β)

Beta denotes the extent up to which an investment will be affected by change in a particular
economic variable. For example, if there is a change in foreign currency rate, not every investment
will feel same impact. Companies will high long foreign currency exposure will feet a greater
impact of depreciation of home currency compared to companies with lower foreign currency
exposure. Companies with earnings in foreign currency will feel a positive impact of depreciation
of home currency.

Calculation of expected return under APT:

Expected returns are calculated as risk free rate plus risk premium of risk factors affecting the
company. APT states, that the expected rate of return of security is the linear function from the
complex economic factors common to all securities and can be estimated using formula:

E (RA) = ƛ0 + β1 ƛ1 + β2 ƛ2 + β3 ƛ3 + β4 ƛ4 + ......... + βn ƛn + εA

Where,

E (RA) = the expected return on stock A

ƛ0 = Risk free rate of return (expected rate of return for the security if the influence of all factors is
0).

ƛ1, ƛ2 , ƛ3 , ƛ4 ......... ƛn = risk premium for each risk factor or macro-economic variable.

β1 , β2 , β3 , β4 , ......... βn = coefficient Beta, showing sensitivity of security’s J rate of return upon

εA = error of rounding for the security (expected value = 0).

It is important to note that the arbitrage in the APT is only approximate.

Risk Factors in the APT

There could be an infinitive number of factors, although the empirical research done by S.Ross
together with R. Roll (1984) identified four factors – economic variables, to which assets having
even the same CAPM Beta, are differently sensitive:

• inflation;

• industrial production;

• risk premiums;

• slope of the term structure in interest rates.

Page No 6 of 8
 Arbitrage Pricing Theory
In practice an investor can choose the macroeconomic factors which seems important and related
with the expected returns of the particular asset. The examples of possible macroeconomic factors
which could be included in using APT model:

• GDP growth;

• an interest rate;

• an exchange rate;

• a default spread on corporate bonds, etc.

Including more factors in APT model seems logical. The institutional investors and analysts closely
watch macroeconomic statistics such as the money supply, inflation, interest rates, unemployment,
changes in GDP, political events, and many others. Reason for this might be their belief that added
information about the changes in these macroeconomic indicators will influence future asset price
movements. But it is important to point out that not all investors or analysts are concerned with the
same set of economic information and they differently assess the importance of various
macroeconomic factors to the assets they have invested already or are going to invest.

At the same time, the large number of the factors in the APT model would be impractical, because
the models seldom are 100 percent correct and the asset prices are function of both macroeconomic
factors and noise. The noise is coming from minor factors, with a little influence to the result –
expected rate of return.

The APT does not require identification of the market portfolio, but it does require the specification
of the relevant macroeconomic factors. Much of the current empirical APT research are focused on
identification of these factors and the determination of the factors’ Betas. And this problem is still
unsolved. Although more than two decades have passed since S. Ross introduced APT model, it has
yet to reach the practical application stage.

The CAPM and APT are not different, because they are developed for deciding an expected rate of
return based on one factor (market portfolio – CAPM) or several macroeconomic factors (APT).
But both models predict how the return on asset will result from factor sensitivities and this is of
immense importance to the investor.

Page No 7 of 8
 Arbitrage Pricing Theory
CAPM V/S APT

CAPM APT

1 It is a single factor model It is a multi-factor model

2 Dependent on market returns. Does not depend on market return.

3 Assumes markets are always correctly priced. States that markets are not always correctly
priced.

4 Assumes borrowing and lending at risk free Does not assume borrowing and lending at
rates. risk free rates.

5 Based on covariance between security returns Based on covariance between security returns
market factors. and many external factors including the
market.

6 Is a special case of APT. Includes CAPM as a special case where a

7 CAPM is suitable only for single period APT works better for multi period returns
returns. evaluation.

8 Assumes that returns follow a normal Does not assume any assumption about
distribution. returns.

9 Makes too many assumptions, hence very Makes fewer assumptions, hence more
restrictive in practical world. realistic in practical world.

10 Data requirements are very less, therefore Requires more data inputs, therefore more
easy to compute. complicated and difficult to compute.

11 User need not find risk factors affecting the User has to find risk factors affecting the
stocks stocks and the extent of their effect on the
stock prices.

12 Risk free rate of return means rate of return Risk free rate of return means return earned if
offered by the Treasury Bonds. there is no change in the level of risk factors
than already assumed.

Page No 8 of 8