You are on page 1of 4

\,ll'ltl'-!;ll'!{). ~ v... t .s. ,,-t .

L ~

'X'l'"ll';:;')""!':: nJ"tl~I;' '- {'\ j '\..


\,... ~ "-'. ._., { '\ 1..)'" ,

<

,_

1)\,1 t

i.

Stephen M Schaefer
London Business School

'"-.

__.,------_._ .._-- ..

__ ._-_._----------------~,-~----

More general versions ofthe CAPM predict risk premia that have \:.'-\<Itt~ the same form as the APT:

-~-------

where the fJ's are the sensitivities of returns on asset i to factors F" F2 FJ etc. and the A's are the factor risk premia

Where do these models come from?

"

Startil1f.: poiflt

is the the states model: relation between risk premia and the stochastic discount factors (sdj's), ms:
w h ere: m = --~-

~-------

A-D price prob(state s)

-;

f(ey isstl/!: what determines the the sdfs:


The challenge is to find the right measure of "bad times," [Le ..) rises in [the stochastic discount factor), so that we can understand high average returns or low prices as compensation for assets' tendency to pay off poorly in "bad times," (John Cochrane, Financial Markels and lhe Real Economy)

Ifwe understand this then we can understand risk premia (cost of capital)

Multi-factor cxtcnsions ofthc CAPM

The stoc/tastic discountfactor (sdj) is a measure of the ,'careitr of available funds and will be relatively ~;"i\ in states when:
I
,'

people have is low


III,'/"

10'1'

'I'{llllt:

Le., value of market portfolio


dri,','
;1/,'(li1lt' ./oltl/:

ir<

I/,//\WI/C'\

Ji/,II

".g., bad

,'

job market (recession) or changes in oPPoJ'tllllities/r people want to save more

investnlent

that make

Each ofthese willlead people to consume less now

Multi-factor extcnsions ofthe CAPM

________________

r.
In CAPM the value of the market portfolio represents total wealth in the economy Investors care only about wealth (because there is only one period and so they don't worry about investing for later periods) The s4{therefore depends onlv on wealth (the value ofthe market portfolio) Qne "route" to CAPM: linear relation between sdfand return on market portfolio
m =a-bR b>O s M s
Multi-factor cxtensions orlhe CAPM

Basics: consumption intluences

affected by both

H\',Ifl

and

1I01'-'l','.;ltll

if 11n",I<
../ '" caJth is lo\\'

, d<
./ othcr factors make investol's save more (311d consume
less)

then consumption will be low and sd{will be high (& vice versa) The \,It should at each point retlect a level of consumption that retlects the best trade-off between these two effects ,!llili\';I!I"II: consumption should be a good proxy for sdf consumption low => sdfhigh consumption high => s4{low
(,

\ilulti-factor extclIsions oflhe CAPM

As with the conventional CAPM, one route to the consumption CAPM is via a linear relation between the sdf and (aggregate) consumption Following the same steps as before we nd:

Aggregate Consumption

',11/ '.

1I/l/i{/1I1l

1\.1ulti.f~ctor cxtensions ofthe CAPM

) lil,~/C

F[[c{(Jr

1.. \ )l1';U

111pl I1 H: \

lmil:!

i:.; (

',11h i <t;:n1

\\ ilh ,HlJi~t{cf()r
,~ '.!.l_~.!.!!.ili!..!.0_~.r::l:
,'

l'();;t (, )rC~q)il~d

:-'1ludcl

explains risk premia in terms of an asset's beta on


(C(I I/\'II I11prio 11)

(me jClIIr

./

just as

rs!!\l.tr

Lv..:.':J.

explains risk premia in terms ofan asset's

beta on olle./c/or (the I//arke/) Level of consumption should be a good measure of"bad times" har: ./
.

When we implement this we have to assume sorne particular relationship between the sdf and the level of consumption .
Lmll1ple: This may not be linear and using other variables to explain sensitivity ofthe sdfto consumption leads to a multi-factor

model. Implies that, although consumption CAPM is a single factor model, it is potentially consistent with a 1l12u,ti-htl<l hl'llllli:lIl"'1
~vt\llti-factor cxtensiol1s orthe CAPM

You might also like