Problem
You are given the following information:
Assume an indirect function ( g)=Eu( g+ y ) .
where E is the expectation operator, g is a random variable, v is a function.
−Eu' (g+ y )
Assume also that Q w ( g )=
Eu' ' ( g+ y )
E u' ( g+ y ) E u' ' ' ( g+ y )
Differentiating Qw ( g ) around g yields Q w ' ( g )= 2
−1
[ E u' ' ( g + y ) ]
D (g)
Differentiating Q w ' ( g ) around g yields Q w ' ' ( g )= 3
[ E u' ' ( g+ y ) ]
Where
D ( g )=[E u( g+ y)' ' ]2 E u' ' ' ( g+ y ) + E u' ( g+ y ) E u' ' ( g+ y ) E u' ' ' ' ( g+ y )−2 E u' ( g+ y )[E u '' ' (g + y)] 2
Now assume that the random variable y is distributed as k ~
x (~
x is another random variable),
~ ~
which has E x=0 and Var ( x )=2.
A 2nd order Taylor expansion around g will give
E u( n) ( g+ k ~
x ) =u(n ) ( g ) +k 2 u( n+2) ( g ) +o( k 3 ),
where (n) is the number of differentiations and o (.) is Landau’s symbol.
Given the above 2nd order Taylor expansion around g,
then D ( g ) would become
2
D ( g )=[ u( 2) ( g )+ k 2 u (4 ) (g) ] [ u( 3) ( g )+ k (2 ) u( 5) ( g) ]+ ( u(1 ) ( g ) +k 2 u( 3) ( g ) u( 2) ( g )+ k 2 u (4 ) ( g ) ) [ u( 4 ) ( g ) +k 2 u( 6) ( g ) ]−2 [ u(1 ) ( g ) +k 2 u( 3 ) ( g
Questions to be answered:
If we have function w ( g) transforming to the following w (g)=Eu( g−eg+ ỹ ), where e is a constant
number and
−Eu' ( g−eg+ y )
Q w ( g )=
Eu' ' (g−eg+ y )
D ( g )=( 1−e )2 [ E u( g−eg+ y )' ' ]2 E u' ' ' ( g−eg + y )+ E u' ( g−eg+ y ) E u' ' ( g−eg+ y ) E u' ' ' ' ( g−eg+ y )−2 E u' ( g−eg+ y
1) How would the 2nd order Taylor expansion would change?
E u( n) ( g−eg+ k ~x ) =?
2) Given the new Taylor expansion (from question 2) how D ( g ) becomes now?