Scribd
Upload
49 views

Factor Model Notes:: R α β f β f …+β f ϵ R α β ' f ϵ

There are three types of factor models: 1. Macroeconomic factor models use observable economic and financial time series data as factors. 2. Fundamental factor models use observable asset characteristics as factors. 3. Statistical factor models extract unobservable factors from asset returns. Factor models express an asset's return as a linear combination of common factors and an idiosyncratic error term. The Sharpe single factor model uses the market excess return as the single factor. The Fama-French 3-factor model builds on this to include additional factors.

Uploaded by

Mohamed Hussien
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
49 views

Factor Model Notes:: R α β f β f …+β f ϵ R α β ' f ϵ

There are three types of factor models: 1. Macroeconomic factor models use observable economic and financial time series data as factors. 2. Fundamental factor models use observable asset characteristics as factors. 3. Statistical factor models extract unobservable factors from asset returns. Factor models express an asset's return as a linear combination of common factors and an idiosyncratic error term. The Sharpe single factor model uses the market excess return as the single factor. The Fama-French 3-factor model builds on this to include additional factors.

Uploaded by

Mohamed Hussien
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 2

Factor Model Notes:

BRIEFING:
Three types of Factor Models:
1. Macroeconomic Factor Models
a. Observable Factors
b. Derived from economics and financial time series data
2. Fundamental Factor Models
a. Observable Factors
b. Derived from asset characteristics
3. Statistical Factor Models
a. Unobservable Factors
b. Extracted from asset returns

General Form:
Rit =α i + β 1 i f 1 t + β 2 i f 12t + …+ β Ki f Kt + ϵ it
Rit =α i + β ' i f t + ϵ it

 Rit is the simple or real return on asset i (i = 1,…, N) in time period t (t = 1,…,T)
 Fkt is the kth common factor (k = 1,…,K)
 Beta is the factor beta for asset I on the kth factor
 Epsilon is the asset specific factor (noise)

Assumptions:
1. The factor realizations, f, are stationary with unconditional moments
a. E [ f ]=μ
b. cov ( f , ϵ )=E [ ( f −μ ) ( f −μ )' ]=Ω
c. Why is this important?
2. Asset specific error terms are uncorrelated with each of the common factors
a. cov ( f , ϵ )=0
3. Error terms are serially uncorrelated and uncorrelated across assets at the same
time frame
2
a. cov ( ϵ it , ϵ js )=σ for all i=j and t=s
b. cov ( ϵ it , ϵ js )=0 otherwise
c. This is because the noise found at one time frame should not play a role on
the noise term in another time frame. This isn’t necessarily true in the real
world.

MACROECONOMIC FACTOR MODELS:


Rit =α i + β ' i f t + ϵ it
 f = observed economic/financial time series factors
 Econometric Problems:
o Choice of Factors
o Estimation of factor betas and residual variances using time series
regression
o Estimate factor covariance matrix from observed history of factors

Sharpe Single Factor Model:


 Sharpe’s single factor model is a macroeconomic factor model that considers one
factor to determine an asset’s return: the market excess return (over risk-free)
 f t=R M t , K=1

 Rit =α i + β i R Mt + ϵ it

 Since the market excess returns are observable, use time series regression to
estimate the parameters beta and variance of the asset
 EXPERIMENT:
o Let’s look at monthly returns, so each time step is 1 quarter
o Let’s look at various stocks in the same industry, S&P Telecom Select
Industry Index

Fama-French 3-Factor Model:

You might also like