Scribd
Upload
6 views

Macro_2017_4

Uploaded by

minthet2023
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
6 views

Macro_2017_4

Uploaded by

minthet2023
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Macroeconomics

Lecture 4

1 Arayah Preechametta 10/15/2018


Stochastic Control Problem
∞
Max E0 ∑ β t r ( xt , ut ), 0 < β <1 (17 )
t =0

s.t. xt +1 = g ( xt , ut , ε t +1 ) , x0 given (18)


ε t +1 is a sequence of i.i.d .(independently and identically distributed )
random var iables.
ε t +1 is realized at t + 1 after ut has been decided at time t.

Bellman ' s equation,

V ( x)
u    (
   )
max  r ( x, u ) + β E V  g x, u , ε  | x   (19 )
The 1st − order condition for the problem on the RHS of (19) is

∂r ( x, u ) (
 ∂g x, u , ε ) 
∂u
+ βE 
 ∂u  (  )
V '  g x, u , ε  | x  =

0 ( 20 )
 
2 Arayah Preechametta 10/15/2018
Stochastic Control Problem

The value function must also satisfies ( this is ana log ous to eq. ( 9 ) )

∂r  x, h ( x )  (
 ∂g x, h ( x ) , ε) 
V '( x)
=
∂x

+ βE
 ∂x
( ( )) 
V ' g x, h ( x ) , ε | x 

( 21)

3 Arayah Preechametta 10/15/2018


Consumption with a Random Return
∞
max E0 ∑ β tU ( ct ) , 0 < β <1
t =0

s.t. At +1 = Rt ( At − ct ) , t ≥ 0, A0 given,
and assumed that U ' ( c ) > 0, U '' ( c ) < 0.
∴ Bellman ' s equation is
V ( A=
t , Rt −1 ) max {U ( At − ut ) + β EtV ( ut Rt , Rt )} ,
ut

where u=
t At − ct ,
The 1st − order condition ( w.r.t. ut )
 ∂V ( ut Rt , Rt ) 
−U '(ct ) + β Et  .Rt | At , Rt −1  = 0. ( 22 )
 ∂ut Rt 

4 Arayah Preechametta 10/15/2018


Consumption with a Random Return
Then from (21),
∂U  At − h ( At , Rt −1 )  ∂ct
V ' ( At , Rt −1 ) = .
∂ct ∂At
 ∂h ( At , Rt −1 ) .Rt 
+ βE V ' ( h ( At , Rt −1 ) .Rt , Rt ) | At , Rt −1 
 ∂At 
∂U [ ct ]  ∂h ( At , Rt −1 ) 
= 1 − 
∂ct  ∂At 
∂h ( At , Rt −1 )
+
∂At  { }
 β E RtV ' ( h ( At , Rt −1 ) .Rt , Rt ) | At , Rt −1 

∂U ( ct ) ∂h ( At , Rt −1 )  ∂U ( ct ) 
=
∂ct
−
∂At
  {
−  β E Rt .V ' ( h ( At , Rt −1 ) .Rt , Rt ) | At , Rt −1 } 

 ∂ct 
∂U ( ct )
= . (1st − order condition w.r.t. ut )
∂ct
Hence, V ' ( At +1 , Rt ) = U ' ( ct +1 ) .
Substituting the above equation int o equation ( 22 ) , one has
U ' ( ct ) = β Et U ' ( ct +1 ) .Rt  ( 23)
⇒ an optimal saving policy function ut = h ( At , Rt −1 )
5 10/15/2018
Let U ( ct ) = ln ct ,
Rt be an i.i.d . random process such that 1 ≤ ERt < (1/ β 2 )
If ct = γ At , then (23) becomes
 Rt   
(= Rt ( At − ct ) )
1 Rt
= β Et  =  β Et  , A
γ At γ
 t +1 
A γ R
 t t ( A − γ A )
t 
t +1

Since Rt is an iid random var iable, hence Et { Rt } = E {Rt } , hence


1  1 
= βE  ,
γ At  γ ( At − γ A )
t 

γ ( At − γ At ) =
βγ At ,
(1 − β ) γ At =
γ 2 At ,
∴ γ=
1− β ,
 =A1 R0 [ A0 − c0 ]
∴ A1 =R0 (1 − γ ) A0 , ( ct =γ At )
A=
2 R1 ( β ) A=
1 R1β R0 (1 − γ ) A=
0 β 2 R1 R0 A0
:
 t −1 
At β= ∏ R j  A0 ,
t
t 1, 2,.....
 j =0 
 t −1 
∴ γ At =−
ct = (1 β ) β ∏ R j  A0 ,
t
t=
1, 2,.....
6  j =0  10/15/2018

You might also like