APT model & Factor model

[florinpop]

is it correct to say that apt model is trying to find the intercept for the factor model?

 

[Black Swan]

Well yes, although it’s not the sole purpose of the APT, the APT was meant originally to be standalone alternative to CAPM (better alternative) although incorprated with a macro surprise model, I believe it to be more effective. Alternatively, they never say it, but theoritically since APT and CAPM should be seeking the same theoretical rate, shouldn’t you also be able to insert CAPM as the intercept in the Macro model even tho it may not be common practice?

 

[florinpop]

thanks

 

[florinpop]

ok I got one more
the factor models have the intercept the expected return… so you only include in other factors surprises… that is if everything is as expected actual return= intercept
while for APT you just take risk free(intercept) and you keep on adding risk premiums (but those risk premiums are value*sensitivity not surprise in value*sensitivity)???

 

[budfox427]

right

 

[florinpop]

cool thanks

 

[pooface]

but i thought the APT is much less reliable than CAPM

 

[mwvt9]

florinpop Wrote:
——————————————————-
> is it correct to say that apt model is trying to
> find the intercept for the factor model?
This only holds for the macro model not the one that uses P/E and stuff (fundamental model?). The intercept on that model is the return for stocks with average exposure to each factor in the model.

 

[Black Swan]

pooface Wrote:
——————————————————-
> but i thought the APT is much less reliable than
> CAPM
I disagree, it’s harder to build in terms of requiring extensive research and somewhat of an art form, but I like the fact that it incorporates several different influences as long as you can minimize multi-collinearlity.

 

[chadtap]

My understanding of the CAPM is for valuing assets which will be held in a complete diverse portfolio… As your only measure of risk is the beta…
The APT model is subject to various influnces/risk measures and would be used for a less deverse portfolio (pricing)…
So regarding Marcofactor models I would think you would only want to use APT models (for intercept) as the assets have various sensitivites to various measures and will correlate with the surprises of the Macrofactor models…
This is just my thought process behind this.. Im not sure if it really matters for exam day..

 

[mwvt9]

CAPM doesn’t hold in the real world. APT can be designed any way you want to…..with whatever risks you believe are “priced” in the market place.