Portfolio Theory AG 924

Week 4.1 Factor models

 Why look at Factor Models?

❑ As a conceptual model, the CAPM is very useful. However, in practice,

with the data and methods that are available to us to measure market
beta, it is not sufficiently useful to compute required rates of return and
expected returns or to discover mispriced securities.

❑ Multifactor models are useful in this context.

❑ These models introduce uncertainty stemming from multiple sources,
whereas the CAPM, in principle, limits risk to one source – covariance
with the market portfolio.
 Roll critique

❑ Roll (1977) argued that the model can not be tested because the tests
involve a joint hypothesis on the model and the choice of the market
portfolio.
❑ On the other hand, many patterns emerge from empirical studies which
are not explained by the CAPM

▪ expected returns and earnings to price ratio have a positive relation (Basu 1977,1983),
▪ small capitalisations have higher expected returns than big capitalisations (Banz 1981),
▪ there is a positive relation between the level of debt and stock returns (Bhandari 1988)
▪ book to market ratio is considered as an explanatory variable in stock returns (Chan,
Hamao and Lakonishok 1991).
▪ Fama and French 1992 on Japanese and American markets respectively.
Arbitrage Pricing Theory
Stephen Ross (1976)
• The Arbitrage Pricing Theory (APT)
– Results in “security market line” for well diversified
portfolios, similar to CAPM, but with a different
process.

Key assumptions:
1. Security returns can be described by a factor model
2. There are sufficient securities to diversify away
idiosyncratic risk
3. Well-functioning security market do not allow for the
persistence of arbitrage opportunities
 Single-factor model
1. Security returns can be described by a factor
model:
𝑅𝑖 = 𝐸 𝑅𝑖 + 𝛽𝑖 𝐹 + 𝑒𝑖
• 𝐸 𝑅𝑖 : expected excess return on stock 𝑖
• 𝐹: new information (surprise) on common
(macroeconomic) factor
• 𝛽𝑖 : sensitivity of firm to the common factor
• 𝑒𝑖 : firm specific news (surprise)
• 𝜎𝐹2 : variance of the common factor, F.
• 𝜎 2 𝑒𝑖 : variance of the firm specific factor 𝑒𝑖 .
• 𝐸 𝑒𝑖 = 0; 𝐶𝑜𝑣 𝐹, 𝑒𝑖 = 0; for 𝑖 ≠ 𝑗, 𝐶𝑜𝑣 𝑒𝑖 , 𝑒𝑗 = 0.
 Portfolios in Single-factor model
• Portfolio excess return:
𝑅𝑃 = 𝐸 𝑅𝑃 + 𝛽𝑃 𝐹 + 𝑒𝑃
𝛽𝑃 = σ 𝑤𝑖 𝛽𝑖 ; 𝐸 𝑅𝑃 = σ 𝑤𝑖 𝐸 𝑅𝑖 ; 𝑒𝑃 = σ 𝑤𝑖 𝑒𝑖 ;
𝐸 𝑒𝑃 = σ 𝑤𝑖 𝐸 𝑒𝑖 = 0 and 𝐶𝑜𝑣 𝐹, 𝑒𝑃 = 0.
• Portfolio variance = systematic + non-systematic:
𝜎𝑃2 = 𝛽𝑃2 𝜎𝐹2 + 𝜎 2 (𝑒𝑃 )
𝜎𝐹2 : variance of the common factor, F.
𝜎 2 (𝑒𝑃 ): non-systematic risk.
 Well-Diversified Portfolios

2. There are sufficient securities to diversify away

idiosyncratic risk.
• Idiosyncratic risk (Non-systematic risk):
𝜎 2 𝑒𝑃 = σ 𝑤𝑖2 𝜎 2 (𝑒𝑖 )
• A well-diversified portfolio has no (or at most
negligible) idiosyncratic risk: 𝜎 2 𝑒𝑃 = 0.
As 𝐸 𝑒𝑃 = 0 and 𝜎 2 𝑒𝑃 = 0, 𝑒𝑃 = 0.
Therefore return of a well-diversified portfolio is:
𝑅𝑃 = 𝐸 𝑅𝑃 + 𝛽𝑃 𝐹
7
 No-arbitrage condition

3. Well-functioning security market do not allow for

the persistence of arbitrage opportunities.
• Law of One Price (LOP)
– Equivalent assets should have the same price.

• Arbitrage
– If LOP violated, arbitrageurs engage in arbitrage
activity:
• Simultaneously buying the asset where it is cheap
and selling where it is expensive.
• The process continues until the market reaches
the no-arbitrage condition. 8
 Arbitrage opportunity: Non-
zero alpha
• Suppose the excess returns of security 𝑖 has non-zero
alpha, 𝛼𝑖 , for certain.
– Alpha is extra return not from common factor nor security
specific factor, e.g. mispricing.
• There is a portfolio 𝑃, well diversified but still yields a
non-zero alpha, 𝛼𝑃 , for certain.
• Investors set portfolio 𝑃′ s expected return as:
𝐸 𝑅𝑃 = 𝛼𝑃 + 𝛽𝑃 𝐸 𝐹
• Assume there is a well-diversified portfolio 𝑀, and it
represents market factor 𝐹:
𝐸 𝑅𝑃 = 𝛼𝑃 + 𝛽𝑃 𝐸(𝑅𝑀 )
9
 Eliminate the risk: Zero-beta
portfolio
• The only risk to gain 𝛼𝑃 is the systematic risk.
Arbitrageurs eliminate this risk by forming a zero-beta
portfolio.
• Suppose 𝑃 has a positive alpha: 𝛼𝑃 > 0.
• Consider a portfolio 𝑍 consists of 𝑃 and 𝑀 with weights
𝑤𝑃 and 1 − 𝑤𝑃 :
𝑅𝑍 = 𝑤𝑃 𝑅𝑃 + 1 − 𝑤𝑃 𝑅𝑀
= 𝑤𝑃 𝛼𝑃 + 𝛽𝑃 𝑅𝑀 + 1 − 𝑤𝑃 𝑅𝑀
= 𝑤𝑃 𝛼𝑃 + 𝑤𝑃 𝛽𝑃 + 1 − 𝑤𝑃 𝑅𝑀
𝛽𝑍 ≡ 𝑤𝑃 𝛽𝑃 + 1 − 𝑤𝑃
𝛼𝑍 ≡ 𝑤𝑃 𝛼𝑃
10
 Eliminate the risk: Zero-beta
portfolio
• To satisfy 𝛽𝑍 = 0,
𝑤𝑃 𝛽𝑃 + 1 − 𝑤𝑃 = 0.
1 −𝛽𝑃
֜ 𝑤𝑃 = ; 1 − 𝑤𝑃 = .
1−𝛽𝑃 1−𝛽𝑃

• Risk premium (expected excess return) of

𝑍 with zero-beta:
𝐸 𝑅𝑍 = 𝛼𝑍 + 𝛽𝑍 𝐸 𝑅𝑀
= 𝑤𝑃 𝛼𝑃
1
= 𝛼𝑃
1−𝛽𝑃
11
 Executing Arbitrage

1
• 𝑍 offers risk-free non-zero alpha, 𝛼𝑍 = 𝛼 .
1−𝛽𝑃 𝑃
– If 𝛼𝑍 > 0: 𝑅𝑍 > 0; 𝑟𝑍 > 𝑟𝑓 .
– 𝑟𝑍 and 𝑟𝑓 are both risk-free.
• 𝑍 offers better rate at same condition, risk free,
i.e. arbitrage opportunity exists.
– Borrow as much as possible at 𝑟𝑓 and buy 𝑍
as much as possible to yield 𝑟𝑍 .
– If 𝛼𝑍 < 0, sell 𝑍 short as much as possible and
invest in risk free asset and yield 𝑟𝑓 .
12
 Executing Arbitrage
1
𝛼𝑍 = 𝛼𝑃
1 − 𝛽𝑃
1
If 𝛼𝑃 > 0 𝑎𝑛𝑑 𝛽𝑃 < 1: > 0, 𝐸 𝑅𝑍 > 0
1−𝛽𝑃
Borrow at risk-free and Buy 𝑍.
1
If 𝛼𝑃 > 0 𝑎𝑛𝑑 𝛽𝑃 > 1: < 0, 𝐸 𝑅𝑍 < 0
1−𝛽𝑃
Sell 𝑍 short and invest at risk-free.
1
If 𝛼𝑃 < 0 𝑎𝑛𝑑 𝛽𝑃 < 1: > 0, 𝐸 𝑅𝑍 < 0
1−𝛽𝑃
Sell 𝑍 short and invest at risk-free.
1
If 𝛼𝑃 < 0 𝑎𝑛𝑑 𝛽𝑃 > 1: < 0, 𝐸 𝑅𝑍 > 0
1−𝛽𝑃
Borrow at risk-free and Buy 𝑍.

13
 To the no-arbitrage condition

• If there is no arbitrageurs in the market, the

arbitrage opportunity does not disappear.

• But if, even a very small number of, investors

realised the arbitrage opportunity, they will bid
up (down) the price of 𝑍 until the positive
(negative) alpha disappears.
– Assuming a “well-functioning market”, i.e. no
frictions to execute arbitrage.
14
 To the no-arbitrage condition

• Therefore, alpha of any well-diversified portfolio

must be zero.
Hence:
𝐸 𝑅𝑃 = 𝛽𝑃 𝐸(𝑅𝑀 )
𝐸 𝑟𝑃 = 𝑟𝑓 + 𝛽𝑃 𝐸(𝑟𝑀 ) − 𝑟𝑓
• Similar to CAPM’s mean-beta relationship!
– Not for individual securities, but for well-
diversified portfolios.
15
 Arbitrage Pricing Theory (APT)

• Arbitrage Pricing Theory (APT):

𝐸 𝑟𝑖 = 𝑟𝑓 + 𝛽𝑖 𝐸 𝑟𝑀 − 𝑟𝑓
▪ 𝐸 𝑟𝑖 : Expected return of well-diversified portfolio 𝑖
▪ 𝑟𝑓 : Risk free rate
▪ 𝛽𝑖 : Sensitivity to the common factor (beta)
▪ 𝐸 𝑟𝑀 : Expected return of Market portfolio (well-
diversified portfolio representing the common market
factor), e.g. FTSE ALL index

16
APT’s Mean-Beta relationship

E(r) 𝐸 𝑟𝑖 = 𝑟𝑓 + 𝛽𝑖 𝐸 𝑟𝑀 − 𝑟𝑓

𝐸(r𝐶 ) C
M
𝐸(r𝑀 )
B
𝐸(r𝐵 )
A
𝐸(r𝐴 )

r𝑓
𝐸 𝑟𝑀 − 𝑟𝑓
β𝐴 β𝐵 β𝑀 β𝐶 β
=1
A, B, and C are well-diversified portfolios
APT vs CAPM

• APT only requires small number of sophisticated

arbitrageurs
– CAPM requires investors are all mean-
variance optimizers
• APT is based on observable portfolios such as
the market index.
– CAPM is not testable in strict sense, as it
relies on unobserved, all-inclusive portfolio.
• APT is silent for individual assets.
 Multifactor Models
❑ Different stocks have different sensitivities to the sources of
systematic risk.
❑Market return captures only average effect
❑Consider the following two-factor model:
Rt =  + GDPGDPt + IRInterest-Ratet + t
Consider two firms and potential effects of the two factors on
their returns:
Regulated utility: low (+) GDP , high (-) IR
Airline: high (+) GDP, low (-) IR
Single factor model cannot capture the effects of both.
 Multi-factor APT

• Multi-factor model:
𝑅𝑖 = 𝐸 𝑅𝑖 + 𝛽𝑖,1 𝐹1 + 𝛽𝑖,2 𝐹2 + ⋯ + 𝛽𝑖,𝑛 𝐹𝑛 + 𝑒𝑖
– 𝐹 are common factors.
• APT requires portfolios representing these
common factors, the factor portfolios.
• Factor portfolio (tracking portfolio):
– Well-diversified portfolio constructed to have a beta
of 1 on a specific factor and 0 on any other factors.
– Portfolio that track the evolution of particular common
factor but uncorrelated with others.
Features of a Multifactor Model
• By definition:
Residual variance of stock i = σ2(ei), where i = 1, ….,
N Variance of factor j = σ2(Fj), where j = 1, …., L

• By construction:
Mean of ei = 0 for all stocks
Covariance between factors j and k = 0
Covariance between residual of stock i and factor j = 0.

• By assumption:
Covariance between ei, ej = 0, for all stocks (i ≠ j).
 Typical Multifactor Models

Chen, Roll and Ross (1986):

Rit =  +iIPIP + iEIEI + iUIUI + iCGCG + iGBGB + it
Where:
- IP: % in industrial production

- EI: % in expected inflation

- UI: % in unanticipated inflation

- CG: Excess return of long-term corporate bonds over long-term

government bonds

- GB: Excess return of long-term bonds over T-bills

Fama-French Three-Factor Model

Fama and French (1996)

𝑅𝑖𝑡 = 𝛼𝑖 + 𝛽𝑖𝑀 𝑅𝑀𝑡 + 𝛽𝑖𝑆𝑀𝐵 𝑆𝑀𝐵𝑡 + 𝛽𝑖𝐻𝑀𝐿 𝐻𝑀𝐿𝑡 + 𝑒𝑖𝑡

– 𝑆𝑀𝐵 [Small Minus Big]:

the return of a portfolio of small stocks in excess of
the returns on a portfolio of large stocks.
– 𝐻𝑀𝐿 [High Minus Low]:
the return of a portfolio of a stocks with a high book-
to-market ratio in excess of the return on a portfolio
of stocks with a low book-to-market ratio
Fama-French Three-Factor Model

• Size and book-to-market ratios explain returns

on securities.
– Smaller firms experience higher returns.
– High book to market firms experience higher
returns (value style v. growth style).
• Returns are explained by size, book to market
and by beta.
– Size and value are priced risk factors,
consistent with APT.
– Alternatively, premiums could be due to
investor irrationality or behavioural biases.
Small Size Effect
• Size effect (Small firm effect)
– Portfolios consist of firms with small market
capitalisation (small size) consistently have
provided higher returns than portfolios of
large size even after the risk adjusted.

• Rolf W. Banz (1981)

– The relationship between return and market
value of common stocks (1981, Journal of
Financial Economics)
Avg. Annual Return for Size-Based
Portfolios of US stocks, 1926 – 2015
Small Size Effect
• Banz 1981
– It is not known whether size per se is
responsible for the effect or whether size is
just a proxy for one or more true unknown
factors correlated with size.

• Potential explanations:
▪ Neglected Firm Effect
▪ Liquidity Effects
Potential explanations
• Neglected Firm Effect:
– Small and less-analysed stocks provide
abnormal risk adjusted returns
– Risk associated to limited information

• Liquidity risk premium:

– Investors hesitate to hold less-liquid (illiquid) asset.
– Small stocks tend to be illiquid.
– Liquidity:
• Ease of sale
• Trading costs
Criticisms against Small Size Effect
• Criticisms:
– Not robust across time
• The effect disappeared since mid-80’s.
– Concentrated in the extremes
• Most of the returns come from very small firms,
so-called “Microcap” stocks.
– January effect
• The size effect is almost entirely due to higher
returns on small stocks in January.
January Effect
• The monthly size effect in the US

Figure from: van Dijk, “Is size dead? A review of the size effect in equity returns”
(2011) Journal of Banking and Finance
Small Size Effect
• Revival of the Size effect
– Asness, Frazzini, Israel, Moskowitz, Pedersen. “Size
matters, if you control your junk” (2018) Journal of
Financial Economics
• Size matters, and in a much bigger way than
previously thought, after controlling for quality or
junk.
– the variability of the size effect is largely due to the volatile
performance of small, low quality junky firms.
– a much stronger and stable size premium emerges that is
robust across time, monotonic in size and not concentrated in
the extremes, prevalent across months of the year.
Value investing and Value effect
• Value investing
– Graham and Dodd “Security Analysis” 1934.
– Identify firms with low prices relative to their
fundamental value.
• Book-to-Market Ratios
– Ratio of the book value of equity to the market value
of equity.
– Book value is a simple measure for the fundamental
value of the firm
• Value effect:
– High Book-to-Market ratio (B/M) predicts high returns.
Avg. Annual Return for B/M based
portfolio of US stocks, 1926–2015
Risks of Value firms
Potential explanations: Value firms are risky
• B/M seems to predict GDP growth and is related
to the business cycle.
• Investment Risk (Investment irreversibility)
Zhang, 2005 “The Value premium” (Journal of Finance)
– Value firms tend to have greater amounts of
tangible capital
– less flexible and less quick in responding to
shocks.
• High and asymmetric adjustment costs
Momentum effect
• Momentum effect
Jegadeesh and Titman, 1993. “Returns to Buying
Winners and Selling Losers: Implications for Stock
Market Efficiency” (Journal of Finance)
▪ Good or bad recent performance of particular
stocks continues over time.
▪ Performance of individual stocks is highly
unpredictable, but portfolios of the best-
performing stocks in the recent past
outperform others.
WML: A Fourth Factor
• Winners minus Losers (WML)
Carhart, 1997 “On Persistence in Mutual Fund
Performance” (Journal of Finance)
𝑅𝑖𝑡 = 𝛼𝑖 + 𝛽𝑖𝑀 𝑅𝑀𝑡 + 𝛽𝑖𝑆𝑀𝐵 𝑆𝑀𝐵𝑡 + 𝛽𝑖𝐻𝑀𝐿 𝐻𝑀𝐿𝑡
𝛽𝑖𝑊𝑀𝐿 𝑊𝑀𝐿𝑡 + 𝑒𝑖𝑡
• The original Fama-French model augmented
with a momentum factor
• It has become a common four-factor model used
to evaluate abnormal performance of a stock
portfolio.
Fama-French Five Factor model
• Fama French three factor model + RMW and
CMA
Fama and French, 2015. “A five-factor asset pricing
model” (Journal of Financial Economics)
– Robust minus Weak (RMW): robust and weak
profitability
– Conservative minus Aggressive (CMA):
conservative and aggressive in investment
pattern
 Performance of Multifactor
Models in Portfolio Analysis
❑ Complexity of MFM increases with the number of factors in the model.
❑ Model’s superiority can be tested on the economic and statistical
significance of the variations in actual and predicted returns.
❑ Cohen and Pogue (1967) report that SIM has more desirable properties
than MFM.
❑ Elton and Gruber (1971):
❑ MFM explains historical correlation matrix better but is poor on predictions.
❑ SIM is superior to MFM in terms of performance.
❑ Use of fundamental factors (economy wide) may lead to better
performance.
❑ Roll and Ross (1980) suggest to use at least three factors to capture
historical variance co-variance matrix.
 Performance of Multifactor
Models in Portfolio Analysis
❑Chan et al (1999) find significant benefits in using factor models. A three factor
(market, size, and book-to-market value of equity) model is sufficient to
capture the pairwise return covariances.
❑Brandt (2010) show that different approaches are to select factors in
MFMs. They include:
❑Factors based on economic theory – e.g. CAPM
❑Factors based on empirical evidence – e.g. Chen et al (1986), Fama-French
(1993)
❑Statistical techniques such as Principle components.
❑ However, moving from theory to empirical to statistical the factors become more difficult to interpret.
Portfolio Theory AG 924

Week 4.2 Black Litterman

Learning outcomes

❑ Understand how the Black-

Litterman model translates
theory into practice.
❑ How to adapt Markowitz to
active management.
❑ Justify active management.
❑ How to make forecasts for
portfolio construction.
What is the Black-Litterman Model?

• The Black-Litterman Model is used to determine

optimal asset allocation in a portfolio
• Black-Litterman Model takes the Markowitz
Model one step further
▫ Incorporates an investor’s own views in
determining asset allocations
Two Key Assumptions

• Asset returns are normally distributed

▫ Different distributions could be used, but using
normal is the simplest
• Variance of the prior and the conditional
distributions about the true mean are known
▫ Actual true mean returns are not known
Basic Idea

1. Find implied returns

2. Formulate investor views
3. Determine what the expected returns are
4. Find the asset allocation for the optimal
portfolio
Implied vs. Historical Returns

• Analogous to implied volatility

• CAPM is assumed to be the true price such that
given market data, implied return can be
calculated
• Implied return will not be the same as historical
return
Implied Returns + Investor Views =
Expected Returns

Risk Aversion Market capitalization Uncertainty of

Covariance Matrix Views
Coefficient weights Views
•δ = (E(r) – rf)/σ2 •∑ •wmkt •(P) •(Ω)
•(Q)

Implied return vector

•Π= δ Σ wmkt
Bayesian Theory

• Traditionally, personal views are used for the

prior distribution
• Then observed data is used to generate a
posterior distribution
• The Black-Litterman Model assumes implied
returns as the prior distribution and personal
views alter it
Expected Returns

E(R) = [(τ Σ)-1 + PT ΩP]-1 [(τ Σ)-1 Π + PT ΩQ]

• Assuming there are N-assets in the

portfolio, this formula computes E(R), the
expected new return.
• τ = A scalar number indicating the uncertainty
of the CAPM distribution (0.025-0.05)
Expected Returns: Inputs

Π= δ Σ wmkt

• Π = The equilibrium risk premium over the

risk free rate (Nx1 vector)
• δ = (E(r) – rf)/σ2 , risk aversion coefficient
• Σ = A covariance matrix of the assets (NxN
matrix)
Expected Returns: Inputs

• P = A matrix with investors views; each row a

specific view of the market and each entry of the
row represents the portfolio weights of each
assets (KxN matrix)
• Ω = A diagonal covariance matrix with entries of
the uncertainty within each view (KxK matrix)
• Q = The expected returns of the portfolios from
the views described in matrix P (Kx1 vector)
Breaking down the views

• Asset A has an absolute return of 5%

• Asset B will outperform Asset C by 1%
• Omega is the covariance matrix

P= Q= Ω=
From expected returns to weights

E(R) = [(τ Σ)-1 + PT ΩP]-1 [(τ Σ)-1 Π + PT ΩQ]

• Expected returns

M = [(τ Σ)-1 + PT ΩP]-1

• Uncertainty of returns

Σp = Σ + M

• New covariance matrix

w = (δ Σp )-1 Π

• Weights
Example 1

• Using Black-Litterman model to determine asset

allocation of 12 sectors
▫ View: Energy Sector will outperform
Manufacturing by 10% with a variance of .025^2
 67% of the time, Energy will outperform Manufacturing
by 7.5 to 12.5%
Complications

• Assets by sectors
▫ We did not observe major differences between BL
asset allocation given a view and market
equilibrium weights
▫ Inconsistent model was difficult to analyze
 There should have been an increase in weight of
Energy and decrease in Manufacturing
Example 2
Model in Practice

• Example illustrated in Goldman Sachs paper

• Determine weights for countries
▫ View: Germany will outperform the rest of Europe
by 5%
Black-Litterman Model
Portfolio Asset Allocation Expected Returns
Advantages and Disadvantages

• Advantages
▫ Investor’s can insert their view
▫ Control over the confidence level of views
▫ More intuitive interpretation, less extreme shifts in
portfolio weights

• Disadvantages
▫ Black-Litterman model does not give the best possible
portfolio, merely the best portfolio given the views
stated
▫ As with any model, sensitive to assumptions
 Model assumes that views are independent of each other
Bibliography

Black, F. and Litterman, R. (1991). “Global Asset Allocation with Equities, Bonds,
and Currencies.” Fixed Income Research, Goldman, Sachs & Company, October.
He, G. and Litterman, R. (1999). “The Intuition Behind Black-Litterman Model
Portfolios.” Investment Management Research, Goldman, Sachs & Company,
December.
Black, Fischer and Robert Litterman, “Asset Allocation: Combining Investor Views
With Market Equilibrium.” Goldman, Sachs & Co., Fixed Income Research,
September 1990.
Idzorek, Thomas M. “A Step-by-Step Guide to the Black-Litterman Model.” Zehyr
Associates, Inc. July, 2004.
Satchell, S. and Scowcroft, A. (2000). “A Demystification of the Black-Litterman
Model: Managing Quantitative and Traditional Construction.” Journal of Asset
Management, September, 138-150.
Portfolio Theory AG 924

Readings: Detecting factors

Detecting factors

❑ Factor exposures are typically determined from time

series regression of stock returns on factor
premiums (factor loadings)
❑ Alternatively, they can be forecast (essential in economic
modelling)
❑ Leading indicators
❑ Stock-market as a discounting mechanism
❑ Running a regression requires sufficient data.
❑ Take overs (weighted average of loadings).
❑ IPO’s – Z scores to find similar rankings.
❑ No data – use industry average
Quantitative portfolio management
Quantitative portfolio management subjects investment
strategies to the rigour of scientific method.
A factor is any variable that may predict stock returns.
Good factors exhibit stable and persistent relationships with
stock returns.
 Economic Factors
• Consumption growth (Real
consumption)
• Corporate bond spread (BBB
versus AAA)
• Industrial Production
• Inflation(CPI or PPI)
• Commodity prices
• Real GDP growth (quarter on
quarter)
• Slope of yield curve (10yr minus 2
year yield)
• Unemployment rate
Creating an economic factor
model
1. Choose the economic variable
2. Determine the Rf rate
3. Define the investment universe
4. Estimate stock returns
5. Collect stock return and factor premum data for each time period
6. Then do a time series regression of the stock return on the factor
premium
7. Calculate each stocks average return as a product of the factor
exposure and the vector containing the stocks factor premium
8. Decompose between diversifiable and non diversifiable return
9. Calculate the correlation between the returns of the stocks in the
investment universe.
Fundamental Valuation
Factors
❑ Dividend yield

❑ Enterprise value to EBITDA

❑ Price to Book

❑ Price to Cash Flow

❑ Price to Earnings
❑ Fama and French (1993)
❑ P/EBITDA

❑ PEG Ratio

❑ Price to Earnings growth to dividend yield

❑ R + D to sales

❑ Price to sales

❑ Size.
 Fundamental solvency and
operating efficiency factors

❑ Cash flow from operations. ❑ Cash conversion cycle (receivables divided

by payables)
❑ Cash ratio (cash and marketable securities
divided by current liabilities) ❑ Cost management index (COGS divided by
operating revenue)
❑ Current ratio (assets divided by liabilities)
❑ Equity turnover (net sales divided by
❑ Quick ratio (cash, accounts receivables and
average shareholders equity)
marketable securities divided by current
liabilities) ❑ Fixed assets turnover (net sales divided by
fixed assets)
❑ Inventory turnover
❑ Receivables turnover
❑ Total Asset turnover
Fundamental operating
profitability factors
❑ Gross profit margin ❑ Return on assets (ROA)
❑ Net Profit Margin ❑ Return on common equity
(ROCE)
❑ Operating profit margin
❑ Return on net assets ❑ Return on owners equity

❑ Return on operating assets ❑ Return on total capital (ROTC)

Financial and liquidity risk
factors
❑ Cash flow coverage ratio ❑ Trading turnover
❑ Debt to equity ratio ❑ Float capitalization
❑ Financial leverage ratio (free float measures
❑ Interest coverage ratio what is available to
❑ Total debt ratio trade)
❑ Number of security
owners (impacts
liquidity)
Technical factors – (unscientific
and not recommended)
❑ Bollinger bands
❑ Channel breakouts
❑ Low price
❑ Momentum
❑ Moving average
❑ Relative strength
❑ Support resistance
Technical factors –(informative)
❑Odd lot balance index (ratio of ❑ Advance decline ratio (advancing
odd lots sales to odd lot securities versus declining)
purchases ❑ Arms index (advance decline
❑ On balance volume ratio divided by upside down
ratio)
❑ Short interest
❑ Upside down ratio (Volume of
❑ Volume
stocks that have up moves with
volume of down moves.
❑ Upside downside volume
 Analyst information
❑ High EPS forecast ❑ Percentage of downward revisions
❑ Percentage of recommendation
❑ High recommendation downgrade's
❑ Low EPS forecast ❑ Percentage of recommendation
upgrade's
❑ Low recommendation
❑ Percent of buys
❑ Median EPS forecast ❑ Percent of sells
❑ Median recommendation ❑ Standard deviation of forecasts

❑ Number of analysts
covering stock
❑ Standardized unanticipated
earnings [EPS(a)-
EPS(e)/SD(EPS)]
Social Responsibility
❑ Community ✓ based on charitable giving
❑ Corporate governance ✓ based on executive compensation
❑ Diversity ✓ Based on minorities
❑ Employee relations ✓ Share of profit
❑ Environment ✓ Spend on clean up
❑ Human Rights ✓ Based on infringements
❑ Product ✓ Based on how it markets it products