Analysis and Comparison of Capital Asset Pricing Model

and Arbitrage Pricing Theory Model

Zhi Li

Queen Mary University of London, School of Mathematics, London, E1 4NS, United Kingdom

[email protected]

Abstract. Establishing an optimal investment risk structure and maximizing re-

turns while bearing less losses have long been an important goal of investors un-
der certain market risk levels. In 1965, William Sharpe developed the capital as-
set price model on the cornerstone of Markowitz's mean-variance model, which
is an idealized description of looking for portfolio loans according to expected
returns and standard deviations. Solved the expected rate of return on a single
asset. Its simple logic and intuitiveness make it easier to measure the relationship
between risk and expected return. Subsequently, Stephen Ross published an arti-
cle entitled "The Arbitrage Theory of Capital Asset Pricing" in the Journal of
Economic Theory in 1976. He combined the factor model with the no-arbitrage
condition to obtain a linear relationship between the expected return and the sys-
tematic risk brought by various macroeconomic variables, that is, the arbitrage
pricing theory. Both CAPM and APT are the core theories discussed in modern
portfolio theory in the dynamically developing capital market, both express the
relationship between expected return and risk, and focus on how to reasonably
price risk. At the same time, both models have certain drawbacks. This article
will gradually analyze the CAPM model and APT, compare the differences be-
tween the two, and summarize their advantages and limitations.

Keywords: Capital asset price model, arbitrage pricing theory, multi-factor

model, comparison, mathematical finance

1 Introduction

1.1 Introduction to CAPM

William Sharpe (1965) developed the capital asset price model based on the mean-
variance model proposed by Markowitz (1952) in Portfolio Selection, which is an ide-
alized description of an asset portfolio that expresses the relationship between expected
return and risk in a linear relationship. The model inherits the assumptions of portfolio
theory: the securities market is efficient, which means the information is completely
symmetric; investors are free to borrow or lend capital at the risk -free interest rate; the
total investment risk is represented by the variance or standard deviation, and the sys-

© The Author(s) 2023

H. Mallick et al. (Eds.): ICEMCI 2022, AEBMR 231, pp. 1899–1906, 2023.
https://doi.org/10.2991/978-94-6463-098-5_213
1900 Z. Li

tematic risk can be represented by the beta coefficient. In addition, investors are re-
quired to be rational and to make investment decisions based on the Markowitz port fo-
lio model; the securities market is friction-less, even without taxes and transaction
costs.[1] Other than that, there are also implicit assumptions: the distribution of returns
on each security is normally distributed; investors can hold any part of a security in a
portfolio. The expression of the CAPM model is:

E (ri ) = rf + βim [E( rm ) − rf ] (1)

E (ri ) represents the expected return of the stock; rf is the risk-free interest rate of the
market; E (rm ) is the expected market return of the market portfolio m; βim is the sys-
tematic risk degree of the asset i in the market m. To represent the level of systematic
risk associated with this asset, the formula βi = Cov(ri , rm )⁄Var (rm ) can be applied,
where βi is the ratio of the covariance of the i-th market portfolio to the variance of the
portfolio return.[2] Then E( rm ) − rf is the market risk premium, The above formula
expresses a straightforward linear connection, which also characterises the CAPM
model's ability to streamline a laborious pricing procedure. It is clear that the asset port-
folio's return is divided into two components: the risk -free interest rate and the risk-
compensation of risky assets. Unsystematic risk is enough to be diversified by a port-
folio with many different assets. Therefore, in this case, the beta coefficient deserves
attention since it may show us the size of the systematic risk of a particular asset in
comparison to the average risk of the market portfolio. So, investors usually need to
calculate the β coefficient to help classify different assets to make investment choices:
when β < 1, it means that the systematic risk of the asset is less than the average risk
of the market portfolio. When β > 1, it means that the systematic risk of the asset is
greater than the average risk of the market portfolio. When β = 1, the asset's systematic
risk is equal to the portfolio's average risk, its return is equal to the market return, or it
will experience the same loss as the market. The vast majority of assets have a beta
coefficient larger than zero, however, the possibility that some assets may have a neg-
ative beta coefficient cannot be completely ruled out, which would mean that their re-
turn would be contrary to the market's average return.[3] According to the research, as
the beta coefficient increases, the degree of compensation required due to non -diversi-
fiable risk is also higher. As a result, high risk must be accompanied by high returns.[4]
CAPM equation can be seen as a relation between assets’ risk premia and their betas,
i.e., security market line
E (ri ) − rf = [ E(rm ) − rf ]The risk premium of an asset is proportional to its beta,
i.e., proportional to its covariance with the market (security market line, SML):
 Analysis and Comparison of Capital Asset Pricing 1901

E(ri)-rf SML

E(rm)-rf

E(rm)-rf M
Figure(a)

βm=1 βi

Fig. 1. Security market line

The ordinate of the coordinate axis is the expected return, and the abscissa is the sys-
tematic risk measure β. The SML is a straight line with the risk-free rate as the intercept
and the market risk premium as the slope. The expected rate of return at point M of the
market portfolio is βm = 1. When the beta value is high, the expected rate of return
from investing in the security is higher; when the beta value is low, the expected rate
of return from investing in the security is lower. In effect, the security market line shows
that an investor's return is proportional to the risk that the investor faces. When the
market is in equilibrium, the risks and returns of any asset or investment portfolio cor-
respond to the securities market line, which means that its actual market price is equal
to the theoretical market equilibrium price. However, in the event of market disequilib-
rium, the points outside the security market line represent the anticipated return -risk
combinations.[5]

1.2 Introduction to APT

Stephen Ross published an article entitled "The Arbitrage Theory of Capital Asset Pric-
ing" in the Journal of Economic Theory in 1976. To achieve a linear relationship be-
tween the expected return and the systematic risk brought on by various macroeco-
nomic factors, he combined the factor model with the no-arbitrage condition, resulting
in the arbitrage pricing theory.[6] The existence of APT is based on three assumptions:
the return of any security can be described by a factor model; there are enough sec urities
in the market to diversify risks; the capital market is in a state of equilibrium, and there
is no arbitrage opportunity.
APT assumes that a security's rate of return is affected by an unknown number of
unknown factors, which are independent of ea ch other. The purpose is to identify these
factors and identify the sensitivity of security returns to changes in these factors. Ac-
cording to the number of factors, it can be divided into one-factor model and multi-
factor model.
The expression for the one-factor model is:

ri = E ( ri ) + βi F + ei (2)
1902 Z. Li

E (ri ) represents the expected rate of return of asset i; F represents the deviation of the
public factor from its expected value; βi is the sensitivity of the security to this factor,
which is the degree of influence of factor F on the rate of return of asset i; ei represents
A company-specific disturbance term that is unpredictable, it is a random variable with
an expected value of 0. Therefore, a theoretical model of one-factor arbitrage pricing
can be built, that is, a portfolio p with a fully diversified unsy stematic disturbance term
ei consisting of n stocks, and its return rate can be expressed as:

rp = E (rp ) + βp F (3)

If F is the market factor m, which is measured by the market risk premium, then when
there is no systematic risk, that is, F=0, the diversified portfolio return should be equal
to the risk-free return:

rp = rf + βp [E(rm ) − rf ] (4)

By inference, it is clear that the expression is the same as the capital asset pricing model
when the market is the only factor. Analyzing APT in an one-factor model can more
clearly observe the relationship between arbitrage and equilibrium, and facilitate the
direct comparison of APT and CAPM.
However, in real economic situations, there are often more than one factor that af-
fects expected returns. Therefore, it will be more realistic and explanatory to analyze
the returns of securities using the multi-factor arbitrage pricing theory. By analysing
two-factor model:

ri = E ( ri ) + βi1 F1 + βi2 F2 + ei (5)

In the two-factor model, F1 and F2 are the two factors that affect the security return,
respectively, and βi1 and βi2 represent the sensitivity to the individual factors of secu-
rity i.
In addition, economic variables such as exchange rate changes, interest rate fluctua-
tions, term premiums, price factors, and industry production growth rates can all be-
come risk factors. From this, a multi-factor model with multiple sources of risk can be
derived:

ri = E ( ri ) + βi1 F1 + βi2 F2 + ⋯ + βin Fn + ei (6)

According to the derivation method of the single-factor arbitrage pricing theoretical

model, the multi-factor arbitrage pricing theoretical model can be obtianed as follow-
ing:

rp = rf + βi1 [ E( rm1 ) − rf ] + βi2 [ E(rm2 ) − rf ] + ⋯ + βin [E(rmn ) − rf ] (7)

 Analysis and Comparison of Capital Asset Pricing 1903

Return(%)

Figur(b)

Fig. 2. Return versus systematic risk

The formula for return versus systematic risk for a well-diversified portfolio can be
represented by the straight line in Figure b. Referring to the assumption made at the
beginning, ATP assumes that there are no arbitrage opportunities in the market, beca use
when there are arbitrage opportunities, every investor in the market seizes the oppor-
tunity to have as many positions as possible to obtain risk -free returns. As more arbi-
trageurs enter the market, this limitless investment size will vanish until there is no
longer a return differential between the two portfolios, at which point the arbitrageur's
trading space will also vanish. Eventually, the price of the security will reach equilib-
rium. As the result, in the market equilibrium, the returns of all fully dispersed portfo-
lios are completely determined by systematic factors, and they will all be on the same
line (security market line) in Figure b, otherwise there will be arbitrage opportunities.

2 The connection between the CAPM model and the APT

model

CAPM and APT are the core theories discussed in modern portfolio theory in the dy-
namically developing capital market, both express the relationship between expected
return and risk, and focus on how to reasonably price risk. Both of them could be ap-
plied to ca pital budgeting, investment performance analysis, and securities valuation.
Systematic risk and unsystematic risk are the two forms of risk that CAPM and APT
identify when it comes to risk classification.[7]
The CAPM model just requires that one risk factor be taken into account. It is a
special case of APT, which is also called the one-factor model. Since market portfolios
are used to calculate security returns, the only variables that might have an impact on
security returns are market risk and macroeconomic conditions. The expected return of
a particular security or portfolio is therefore dependent on the beta coefficient. It can
be said that APT, as a multi-factor model, has a wider scope of application and stronger
1904 Z. Li

practicability.[8] If only one risk factor condition is involved, its specific situation is
consistent with the CAPM model.

3 Comparing CAPM model and APT model

Although both CAPM and APT show the relationship between expected return and risk
in a linear form, they essentially have different modeling thinking angles. Markowitz's
mean-variance model is the foundation of the CAPM, which is the outcome of market
equilibrium under mean-variance preference. It focuses on maximizing returns on the
basis of controlling risks, or avoiding risks to the greatest extent on the basis of con-
trolling returns. In general, CAPM examines how assets are valued when all investors
make comparable investments and the market eventually adjusts to an equilibrium with
static characteristics. APT is based on the theory of no-arbitrage equilibrium, relying
on a multi-factor model, deriving returns from the process of generating stock returns,
and using the concept of arbitrage to describe the formation of equilibrium, which is a
dynamic process. In order to generate risk-free profits, investors create positions as
large as possible through the arbitrage portfolio when there is an arbitrage opportunity
in the market. As this situation continues to evolve, the supply and demand among
securities change accordingly. The APT model focuses on how the asset is valued when
there is no risk-free arbitrage in the final market and it achieves equilibrium.[9] From
non-equilibrium to equilibrium, from the existence of arbitrage opportunities to the pro-
cess of no-arbitrage equilibrium, CAPM depends on a large number of investors to
make small adjustments to their positions. By contrast, APT theoretically only requires
one arbitrageur to maintain the market without arbitrage state because it is a risk -free
arbitrage opportunity.
The assumptions of the CAPM model and the APT model are also different. Com-
pared to the APT model, CAPM's assumptions are very strict, which results in it being
limited to a "single investment period". CAPM requires a portfolio based on an efficient
market to complete the analysis and ignores taxes and transaction costs, assuming that
the market is frictionless. Moreover, in terms of restrictions on investors, all investors
are required to be risk-averse, and to have the same view of the security evaluation and
economic situation, which is called the consensus expectation assumption. Based on
the above assumptions, the APT model does not have these constraints and does not
clearly specify investors' risk appetite, and does not require investors to plan and im-
plement investment strategies within a single investment horizon. Additionally, CAPM
stresses that market portfolios must be an effective portfolio, but APT does not espe-
cially emphases the importance of market portfolios. APT does not analyze a single
investment period, and there is no tax problem. Investors can freely borrow and lend
funds at risk-free interest rates, which is more realistic.[10] Although the CAPM mod-
el's strong fundamental assumptions make the mathematical formulation of the model
easier to understand, these criteria cannot yet be satisfied, even on the assumption that
the securities market is getting increasingly more developed. In contrast, the assump-
 Analysis and Comparison of Capital Asset Pricing 1905

tions and conditions of the arbitrage pricing model are less restrictive, but the mathe-
matical expression has a very significant complexity and is more comprehensive and
adaptable.
In terms of risk interpretation, the CAPM model describes security risk by relyin g
solely on a security's beta coefficient in relation to the market portfolio. Although this
will inform investors of the size of the risk, it will not pay attention to where the risk
comes from. As a single index model, CAPM disregards the influence of factors beyond
the market, believing that only complete market forces influence the return o f stocks.
In contrast to the APT model, it acknowledges that security returns are linearly related
to a group of indices that indicate various factors (such as market factors, inflation,
industry factors, interest rate changes, etc.) [11] that affect stock returns. The APT
model thereby broadens investors' horizons of thought since it gives them an analytical
instrument for locating the origin of security risk as well as the ability to assess risk at
various levels.

4 Conclusion

Through the above analysis and comparison of CAPM model and APT model, it is clear
that both models have some shortcomings. However, both theoretically and practically
play an irreplaceable role in considering the "reasonableness" of different securities
prices. The biggest advantage of CAPM is its simplicity and clarity, so it is more con-
venient to use and is widely used in the calculation of various asset prices (such as
human capital pricing, insurance rate calculation, securities market and real estate in-
vestment, etc.). However, due to a series of strict assumptions in the model, and when
considering systematic risk, only the market portfolio is concerned. This is an example
of the actual economy being simplified and the existence theory being abstract ed, but
it is also lack of comprehensiveness. Therefore, while choosing stocks, investors should
consider both the macroeconomic climate and the company's own development in ad-
dition to the beta coefficient. But CAPM models are still useful when studying t he im-
pact of the overall economy on individual stocks. As opposed to this, the multi-factor
APT model explains the risk of securities using a variety of factors, and the beta coef-
ficients of securities returns to various macro factors vary, which is more a ccurate. As
a result, APT may be seen as a specific instance of CAPM, while CAPM can be seen
as a supplement to and modification of APT. However, the APT model also has certain
drawbacks: it cannot clearly point out the relevant risk factors and risk premium; and
in the calculation process, with the continuous increase of the number of risks, the profit
analysis of arbitrage gradually becomes complicated, and the difficulty of related oper-
ations increase as well. Both models, in a word, embody the core idea s of contemporary
financial theory. The CAPM model and the APT model for venture capital offer only
limited and profound insights to investors doing venture capital operations.
1906 Z. Li

References
1. Sharpe. W. Captial asset prices: a theory of market equilibrium under conditions of risk [J].
Journal of Finance, 1964, 19: 425-442
2. “Times Finance” NO. 02, 2017 (Cumulativety NO. 651), A Comparative Study of Capital
Asset Pricing Model and Arbitrage Pricing Model, Chakaiwen
3. “Academic Exchange” Aug 2017, Capital Asset Pricing Model and Arbitrage A Compara-
tive Study of Pricing Models, Lansha
4. “Public Investment Guide”611731, The application of capital asset pricing model,
Yuanyuan
5. Mullins, D. (1982). Does the Capital Asset Pricing Model Work? [online] Harvard Business
Review. Available at: https://hbr.org/1982/01/does-the-capital-asset-pricing-model-work.
6. Ross S A. The arbitrage theory of capital asset pricing [J]. Journal of Economic Theory,
1976 (13): 341-360
7. “Journal of Chongqing College of Electronic Engineering”, A comparative analysis of the
capital asset pricing model and the arbitrage pricing model,20-2, 03/2012, DingYao
8. “Research on Quantitative Economy and Technology Economics” 2001 No.5, An Empirical
Study of Arbitrage Pricing Theory, Linxin
9. Business Bliss Consultants FZE. (November 2018). Literature Review of Arbitrage Pricing
Theory (APT). Retrieved from https://ukdiss.com/litreview/literature-review-of-arbitrag e-
pricing-theory.php?vref=1
10. “Cooperative Economy and Technology”, Issue 5, 2019, Cheng Jing
11. MBA Knowledge Base. (2014). Arbitrage Pricing Theory (APT). [online] Available at:
https://www.mbaknol.com/investment-management/arbitrage-pricing-theory-apt/.

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-
NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/),
which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any
medium or format, as long as you give appropriate credit to the original author(s) and the source,
provide a link to the Creative Commons license and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter’s Creative
Commons license, unless indicated otherwise in a credit line to the material. If material is not
included in the chapter’s Creative Commons license and your intended use is not permitted by
statutory regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder.