random horizon stochastic dynamic slutsky equation under ... · center of game theory, st...
TRANSCRIPT
Applied Mathematical Sciences, vol. 8, 2014, no. 147, 7311 - 7340
HIKARI Ltd, www.m-hikari.com http://dx.doi.org/10.12988/ams.2014.4112
Random Horizon Stochastic Dynamic Slutsky
Equation under Preference Uncertainty
David W. K. Yeung
Center of Game Theory, St Petersburg State University, Russia and
SRS Consortium for Advanced Study in Cooperative Dynamic Games
Shue Yan University, Hong Kong
Copyright © 2014 David W.K. Yeung. This is an open access article distributed under the
Creative Commons Attribution License, which permits unrestricted use, distribution, and
reproduction in any medium, provided the original work is properly cited.
Abstract
This paper extends Slutsky’s classic work on consumer theory to a random
horizon stochastic dynamic framework in which the consumer has an
inter-temporal planning horizon with uncertainties in future incomes, preferences
and life-span. Utility maximization leading to a set of ordinary wealth-dependent
demand functions is performed. A dual problem is set up to derive the wealth
compensated demand functions. This represents the first time that
wealth-dependent ordinary demand functions and wealth compensated demand
functions are obtained under these uncertainties. The corresponding Roy’s identity
relationships and Slutsky equations in a random horizon stochastic dynamic
framework with uncertain preferences are derived. The analysis incorporates
realistic characteristics in consumer theory and advances the conventional
microeconomic study on consumption to a more realistic optimal control
framework.
Keywords: Optimal consumption, Stochastic dynamic programming, Roy’s
identity, Slutsky equation
1 Introduction
The ground-breaking work by Slutsky (1915) laid the foundation for rigorous
analysis of optimal consumption decision in microeconomics. The analysis, which
7312 David W. K.Yeung
showed that the effect of a price change on the demand of a good can be
decomposed into substitution effects and income effect, yields significant
economic implications (see Varian (1992)). This prominent contribution in
consumer theory, known as the Slutsky equation, was christened by John Hicks as
the ‘Fundamental Equation of Value Theory’. It becomes an integral part of
mainstream economics and consumer theory. The papers by Allen (1936 and
1950), Hicks and Allen (1934), Schultz (1935), Dooley (1983) and Epps (1975)
propagated Slutsky’s classic work. Another milestone in consumer theory is
Roy’s identity (1947) which provides an often invoked mathematical result in
consumer theory. In addition, the identity is also instrumental in proving the
Slutsky equation. Yeung (2013) extended Slutsky’s work to a stochastic dynamic
framework in which the consumer has a T -period life-span with future incomes
being uncertain and derived the stochastic dynamic Slutsky equations. Yeung
(2014) further extended Slutsky’s consumer problem into a dynamic framework in
which the consumer has a random life-span and with uncertain future incomes.
In this paper, uncertainties in the consumer’s future preferences is
incorporated to Yeung’s (2014) extension of the Slutsky framework to reflect an
often observed reality in consumer choice. Changes in conditions that could not
be perfectly foreseen like health, taste, habits, technology, style, culture and
family composition contribute to uncertainty in future preferences. In particular,
optimal consumption choice under a dynamic framework with uncertainties in the
consumer’s life-span, future incomes and future preferences is examined.
Inter-temporal wealth-dependent ordinary demand functions and wealth
compensated demand functions are obtained. Two of the most crucial foundations
in consumer theory – Roy’s identity and Slutsky equation – are derived in a
random horizon stochastic dynamic framework with uncertainty in the consumer’s
future preferences.
The paper is organized as follows. We first present a dynamic model of
utility maximization by a consumer with an uncertain life-span, uncertain future
incomes and uncertain future preferences in Section 2. In Section 3, a set of
wealth-dependent ordinary demands is characterized. The Roy’s identity result in
a random horizon stochastic dynamic framework with uncertain future
preferences is derived in Section 4. The dual problem is formulated in Section 5
and the corresponding wealth compensated demand functions are obtained.
Stochastic dynamic Slutsky equations for the consumer with uncertainties in
life-span, future incomes and preferences are formulated in Section 6. Section 7
concludes the paper.
2 Dynamic Utility Maximization under Uncertainties in Incomes,
Preferences and Life-span
Consider the case of a consumer whose life-span involves T̂ periods where T̂
Random horizon stochastic dynamic Slutsky equation 7313
is a random variable with range },,2,1{ T and corresponding probabilities
},,,{ 21 T . Conditional upon the reaching of period , the probability of the
consumer’s life-span would last up to periods ,1, T, becomes respectively
(see Yeung and Petrosyan (2011)):
T
,
T
1
T
T
,, . (2.1)
We use ),,,( 21 kn
kkkk xxxx to denote the quantities of goods consumed and
),,,( 21 kn
kkkk pppp the corresponding prices in period },,2,1{ Tk . The
preference or utility function of the consumer in period 1 is known to be )( 1
)1(1 xu .
His future preferences are not known with certainty. In particular, his utility
function in period },,3,2{ Tk is known to be )()(
k
kxu k with probability
k
k
for },,2,1{ kk m if he survives in period k . We use k~ to denote the
random variable with range },,2,1{ kk m and corresponding probabilities
},,,{ 21 km
kkk .
The consumer maximizes his expected inter-temporal utility
TE ,,, 32
T
TT
1ˆˆ
)()(
11
ˆ
1
k
kk
c
c
m
k
T
k
xu k
k
k
k
T
E ,,, 32 )( 1
)1(1 xu
T
TT
2ˆˆ
)()(
11
ˆ
2
k
kk
c
c
m
k
T
k
xu k
k
k
k
, (2.2)
subject to the budget constraint characterized by the wealth dynamics
1
11
1 )(
k
n
h
h
k
h
kk
n
h
h
k
h
kkk
kk
xpWrxpWW , 0
11 WW . (2.3)
where r is the interest rate, is the consumer’s subjective one-period
discount factor for the duration from period 1 to period , the initial period
discount factor 11 , 1k is the random income that the consumer will
receive in period 1k ; and k , for },,3,2{ Tk , is a set of statistically
independent random variables, and 122 ,,, T
E is the expectation operation with
7314 David W. K.Yeung
respect to the statistics of 2 , 3 , 1, T . The random variable k has a
non-negative range },,,{ 21 km
kkk with corresponding probabilities
},,,{ 21 km
kkk . The random variable 1T has a value of zero with probability 1
because the consumer will receive no income in period 1T . Moreover, under
the axiom of non-satiation, the consumer will spend all his wealth in the last
period of his expected life-span and therefore 01 TW . The consumer’s discount
factor c may or may not coincide with the market discount factor r1
1 and
may also be equal to 1. For notational simplicity we adopt the notation
k
c
c
k
1
1 and
k
c
c
k
.
The problem (2.2)-(2.3) is a random horizon discrete-time stochastic control
problem with uncertain payoffs (see Yeung and Petrosyan (2013 and 2014)).
Now consider the case when the consumer has lived to period t and his
wealth is W and his preference is )()(
t
txu t . The consumer problem can be
formulized as the maximization of the (discounted) payoff:
121 ,,, TttE )(
)(
1 t
tt xu t
T
t
TT
tT
ˆ
1ˆ
)()(
1
1
ˆ
1
k
kkm
k
T
tk
xu k
k
k
k
, (2.4)
subject to the budget constraint characterized by the wealth dynamics
11 ))(1( kkkkk xpWrW , WWt , (2.5)
for },,1,{ Tttk .
In a stochastic dynamic framework, strategy space with state-dependent
property has to be considered. In particular, given that the preference of the
consumer is )()(
k
kxu k if he survives to period k , a pre-specified class of
mapping )()(k
k
XW : with the property
},,,{ 21 kn
kkk xxx kx )()(Wk
k
)}(,),(),({)(2)(1)(
WWW kkkk n
kkk
, for
},,2,1{ Tk , is the strategy space and each of its elements is a admissible
strategy. We define the value function ),()(
WtV t and the set of strategies
)(*)(* Wx k
kk
, for },,1,{ Tttk which provides an optimal
consumption solution as follows:
Random horizon stochastic dynamic Slutsky equation 7315
),()(
WtV t Ttt xxx ,,, 1
m a x
121 ,,, TttE )(
)(
1 t
tt xu t
T
t
TT
tT
ˆ
1ˆ
)()(
1
1
ˆ
1
k
kkm
k
T
tk
xu k
k
k
k
WWt
121 ,,, Ttt
E )]([)()(
1 Wu tt
t
tt
T
t
TT
tT
ˆ
1ˆ
)]([)()(
1
1
ˆ
1
kk
kkm
k
T
tk
Wu kk
k
k
k
WWt ,
for },,2,1{ Tt . (2.6)
The value function ),()(
WtV t reflects the expected inter-temporal utility (in
present value terms) that the consumer will obtain from period t to the end of his
life-span. Following the analysis of Yeung and Petrosyan (2013 and 2014) one
can derive an optimal solution to the random-horizon consumer problem
(2.2)-(2.3) as follows:
Theorem 2.1. A set of consumption strategies )({*)(
Wk
k
, for
},,2,1{ tt m and }},,2,1{ Tk provides an optimal solution to the
random horizon consumer problem (2.2)-(2.3) if there exist functions
),()(
WtV t , for },,2,1{ Tt and },,2,1{ tt m , such that the following
recursive relations are satisfied:
),1()(
WTV T 0 and 01 TW ,
),()(
WTV T
)])(1(,1[)(max)()(
1 TTT
TT
xxpWrTVxu TT
T
,
),()(
WtV t
txmax
1t
E
)()(
1 t
tt xu t
T
t
T
t
1
1
1
1
1
1
t
t
t
m
t
]))(1(,1[ 1
)( 1
ttt xpWrtV t
,
for t }1,,2,1{ T and },,2,1{ tt m . (2.7)
7316 David W. K.Yeung
Proof. Following Bellman’s (1957) technique of dynamic programming we begin
with the last stage/period. By definition, the utility of the consumer at period
1T is therefore
0),1()(
xTV T .
We first consider the case when the consumer survives in the period T , his
preference is )()(
T
Txu T , and the state WWT . The problem then becomes
]))(1(,1[)(max 1
)()(
11
TTTT
TT
xxpWrTVxuE TT
TT
(2.8)
subject to
11 ))(1( TTTT xpWrW . (2.9)
Since 01 T with probability 1, ),1()(
WTV T 0 and 01 TW , the problem
in (2.8)-(2.9) can be expressed as
),()(
WTV T
)])(1(,1[)(max)()(
1 TTT
TT
xxpWrTVxu TT
T
, (2.10)
subject to
))(1(1 TTT xpWrW , WWT and 01 TW . (2.11)
Now consider the problem when the consumer survives to period 1T and his
preference is governed by )( 1
)(1 1
T
Txu T .The problem in period 1T can be
expressed as maximizing
1, TTE
1
1
T )( 1
)(1 1
T
Txu T
T
T
T
1
T
T
T
m
T
1
T
1 )()(
T
Txu T
(2.12)
subject to
11 ))(1( kkkk xpWrW , for },1{ TTk and WWT 1 . (2.13)
If the value function ),()(
WTV T in (2.10) exists, the problem (2.12)-(2.13)
can be expressed as a single stage problem:
Random horizon stochastic dynamic Slutsky equation 7317
),1()1( WTV T
1
1
,maxTT
T
Ex
1
1
T )( 1
)(1 1
T
Txu T
T
T
T
1
T
T
T
m
T
1
]))(1(,[ 11
)(
TTT xpWrTV T
. (2.14)
Now consider the problem that the consumer survives in period t }2,,2,1{ T
and his preference is )()(
t
txu t . Following the analysis above, the problem in
period t becomes the maximization of the expected payoff
121 ,,, TttE
t
1 )()(
t
txu t
T
t
T
tTT
1ˆ
ˆ
)()(
1
1
ˆ
1
k
kkm
k
T
tk
xu k
k
k
k
121 ,,, Ttt
E t
1 )()(
t
txu t
T
t
T
tTT
T
t
T
t
1
1ˆˆ
1
)()(
1
1
ˆ
1
k
kkm
k
T
tk
xu k
k
k
k
(2.15)
subject to
11 ))(1( kkkk xpWrW , for },,1,{ Tttk and WWt . (2.16)
Note that the term
132 ,,, TttE
T
t
T
tTT
1
1ˆˆ
)()(
1
1
ˆ
1
k
kkm
k
T
tk
xu k
k
k
k
can be expressed as:
132 ,,, TttE
T
t
T
tTT
1
1ˆˆ
)()()(
1
1
ˆ
2
1
)(11
1
1
11
1
1
1
k
kkm
k
T
tk
t
ttm
t xuxu k
k
k
kt
t
t
t
.
7318 David W. K.Yeung
In (2.17) the term (2.17)
132 ,,, TttE
)()()(
1
1
ˆ
2
1
)(11
11
k
kkm
k
T
tk
t
tt xuxu k
k
k
kt
gives the expected inter-temporal utility to be maximized in period 1t if the
consumer’s preference in period 1t is )( 1
)(1 1
t
txu t . If the value function
),1()( 1 WtV t
exists, we have:
),1()( 1 WtV t
Ttt xxx ,,, 11
max
132 ,,, Ttt
E
T
t
T
tTT
1
1ˆˆ
)( 1
)(11
11
t
tt xu t
)()(
1
1
ˆ
2
k
kkm
k
T
tk
xu k
k
k
k
. (2.18)
Using (2.18) the problem (2.15)-(2.16) can be formulated as a single stage
problem which maximizes the expected payoff
1tE )(
)(
1 t
tt xu t
T
t
T
t
1
1
1
1
1
1
t
t
t
m
t
]))(1(,1[ 1
)( 1
ttt xpWrtV t
. (2.19)
If ),()(
WtV t exists, we have
),()(
WtV t
1
maxt
t
Ex
)()(
1 t
tt xu t
T
t
T
t
1
1
1
1
1
)(
1
t
t
t
m
t
]))(1(,1[ 1
)( 1
ttt xpWrtV t
,
for t }2,,2,1{ T and },,2,1{ tt m . (2.20)
Hence Theorem 2.1 follows. ■
The stochastic optimal state trajectory derived from Theorem 2.1 is
characterized by:
Random horizon stochastic dynamic Slutsky equation 7319
0
11 WW ,
12
2 ;
2
j
W 21
2
0
1
)(
11
0
1 )]()[1(j
WpWr ,
213
32
2 ;
3
jj
W )]()[1( 12
2212
2 ;
2
)(
22
;
2
jj
WpWr 3
3
j , if preference is )( 2
)(2 2 xu
in
period 2;
3214
43
32
2 ;
4
jjj
W )]()[1( 213
32
23213
32
2 ;
3
)(
33
;
3
jjjj
WpWr 4
4
j ,
if preference is )( 3
)(3 3 xu
in period 3;
121
33
22 ; T
TjT
jj
TW
)]()[1( 2211
13
32
212211
13
32
2 ;
1
)(
11
;
1
TTj
Tjj
TTTj
Tjj
TTTT WpWr
Tj
T ,
if preference is )( 1
)(1 1
T
Txu T in period 1T ;
TTj
Tjj
TW 21
11
33
22 ;
1
)]()[1( 1213
32
21213
32
2 ;)(; TTj
Tjj
TTTj
Tjj
TTTT WpWr
1
1
Tj
T
0 .
We use *
W to denote the set of possible values of wealth 1213
32
2 ;
jjj
W
at period along the optimal trajectory generated by Theorem 2.1.
3 Wealth-Dependent Ordinary Demand
In this section, we consider the primal problem of deriving wealth-dependent
ordinary demand functions in which the consumer maximizes his inter-temporal
expected utility subject to uncertain inter-temporal budget, life-span and future
preferences. Following the analysis in Yeung (2014) we first consider the case
when the consumer survives in the last period and his preference is )()(
T
Txu T .
Let *0
TT WW denote the consumer’s wealth in period T . Given that
),1()(
WTV T 0 and 01 TW , to exhaust all the wealth in this period,
TTT xpW 0 0 . Hence the consumer faces the problem
)(max)(
T
T
x
xu T
T
subject to TTT xpW 0 0 . (3.1)
Problem (3.1) is a standard single period utility maximization problem.
Setting up the corresponding Lagrange problem and performing the relevant
maximization one obtains a set of first order conditions. It is well-known (see
Cheung and Yeung (1995)) that if the set of first order conditions satisfies the
7320 David W. K.Yeung
implicit function theorem, one can obtain the ordinary demand as explicit
functions of the parameters 0
TW and Tp , that is:
h
TTx
)(),( 0)(
TT
h
T pWT , for },,2,1{ Tnh and },,2,1{ tt m . (3.2)
Note that ),( 0)(
TT
h
T pWT corresponds to the optimal consumption strategies
)( 0)(
T
h
T WT in Theorem 2.1. To show this we note that the problem in period T
in Theorem 5.1 is
]))(1(,1[)(max 1
)()(
11
TTTT
TT
xxpWrTVxuE TT
TT
)(max
)(
1 T
TT
xxu T
T
subject to ))(1(1 TTT xpWrW 0 .
Since the problem )(max)(
T
T
x
xu T
T
subject to TTT xpW 0 0 and the
problem and the problem )(max)(
T
TT
xxu T
T
subject to the same constraint yields
the same controls, therefore ),( 0)(
TT
h
T pWT = )( 0)(
T
h
T WT .
Substituting (3.2) into (3.1) yields the indirect utility function in period T
as ),( 0)(
TT
TpWv T )],([ 0)()(
TTT
TpWu TT . Invoking the definition in (2.6),
T
1 ),( 0)(
TT
TpWv T equals the function ),( 0)(
TWTV T in Theorem 2.1.
Now consider the case when the consumer survives in period 1T , and his
preference is )( 1
)(1 1
T
Txu T . If wealth equals *
1
0
1 TT WW in this period, the
problem in concern becomes
TT xx ,1
max
1, TTE )( 1
)(1 1
T
Txu T
T
T
T
1
T
T
T
m
T
1
T )()(
T
Txu T
subject to the inter-temporal budget
TTTTT xpWrW ))(1( 111 ,
11 ))(1( TTTTT xpWrW 0 , 0
11 TT WW *
TW . (3.3)
Let j
TTTTT xpWrW Tj
T
))(1( 11
0
1
; 1 denote the wealth at period T if j
T },,,{ 21 Tm
TTT occurs and the consumer’s preference is )( 1
)(1 1
T
Txu T in
period 1T . Using the indirect utility function ),( 0)(
TT
TpWv T and (3.3), the
problem facing the consumer in period 1T can be expressed as a single-period
problem:
Random horizon stochastic dynamic Slutsky equation 7321
1
maxTx
)( 1
)(1 1
T
Txu T
T
T
T
1
j
T
m
j
T
1
T
T
T
m
T
1
T
),( 1;)(
TT
TpWv T
jTT
. (3.4)
First order condition for a maximizing solution yields
)( 1
)(1 1
1
T
T
xxu T
iT
T
T
T
1
j
T
m
j
T
1
T
T
T
m
T
1
),( 1;)(
TT
T
WT pWv Tj
TT
T
i
Tp 1 )1( r 0 ,
for },,2,1{ 1 Tni ; (3.5)
Again, with the implicit function holding, (3.5) can be solved to yield the ordinary
demands in period 1T as
h
TTx
)(
11
),,( 1
0
1
)(
11
TTT
h
T ppWT
, for
},,2,1{ 1 Tnh .
Note that ),,( 1
0
1
)(
11
TTT
h
T ppWT
corresponds to the optimal consumption
strategies )( 0
1
)(
11
T
h
T WT in Theorem 2.1. To show this we transform (3.4) into a
similar problem by multiplying the maximand by 1
1
T and obtain
)(max 1
)(11
11
1
T
TT
xxu T
T
T
T
T
1
j
T
m
j
T
1
T
T
T
m
T
1
T
1
),( 1;)(
TT
TpWv T
jTT
.
(3.6)
Moreover, recalling that T
1 ),( 0)(
TT
TpWv T equals the ),( 0)(
TWTV T in
Theorem 2.1 we can express (3.6) as
1
maxTx
T
E
)( 1
)(11
1
T
TT xu t
T
T
T
1
T
T
T
m
T
1
]))(1(,[ 11
)(
TTT xpWrTV T
,
which is the period 1T condition to be maximized in Theorem 2.1. Since the
controls for problem (3.4) and the transformed problem (3.6) are the same, we
have ),,( 1
0
1
)(
11
TTT
h
T ppWT
)( 0
1
)(
11
T
h
T WT .
Substituting ),,( 1
0
1
)(
11
TTTT ppWT
into (3.4) yields the inter-temporal
7322 David W. K.Yeung
indirect utility function ),,( 1
0
1
)(1 1
TTT
TppWv T
. Invoking (2.6) and Theorem 2.1,
the term 1
1
T ),,( 1
0
1
)(1 1
TTT
TppWv T
corresponds to ),1( 0)( 1
TWTV T in
Theorem 2.1.
Repeating the analysis for periods 2T to 1 yields the consumer problem
at period }2,,2,1{ T with preference )()(
xu
as:
)(m a x)(
xux
T
T
1 1
1
1
1
1
j
m
j
1
1
1
1
1
m
1 );(;
1
)(1 111 pWv
j
,
(3.7)
where 0
W *
W , 11
1
1
0;
1 ))(1(
j
xpWrWj
, and );(;
1
11 pW
j
is the
short form for the vector ),,,,( 21
;
1
11
TpppWj
.
First order condition for a maximizing solution to the problems in (3.7) can
be obtained as:
)()(
xu ix
T
T
1 1
1
1
1
1
j
m
j
1
1
1
1
1
m
1 );(;
1
)(1 111
1pWv
j
W
0)1( rpi
,
(3.8)
for },,2,1{ ni and }2,,2,1{ T .
Note also the condition that in period , good i will be consumed up the
point where marginal utility of consumption )()(
xu ix
equals )1( r ip times
the expected marginal utility of wealth
T
T
1 1
1
11
1
1
1
W
jm
j
v
1
1
1
1
1
m
);(;
1
)(1
1
111
1pWv
j
W
.
In particular, the expected marginal utility of wealth takes into consideration
the random future income and preferences plus the probability of the consumer
surviving in period 1 . Solving (3.8) yields the ordinary demands in period
as:
h
x)(
),( 0)(pW
h
, for },,2,1{ nh and },,2,1{ hh m . (3.9)
Random horizon stochastic dynamic Slutsky equation 7323
After solving the primal consumer problem which maximizes expected
utility subject to uncertainties in future income, future preferences and life-span,
we proceed to derive the Roy’s identity result in a dynamic framework with
uncertainties in future life-span, preferences and incomes.
4 Random Horizon Stochastic Dynamic Roy’s Identity under
Uncertain Preferences
In this section we derive the stochastic dynamic version of Roy’s identity with
uncertainties in future life-span, preferences and incomes. Invoking (3.7) we
obtain the identity
),( 0)(pWv
)],([ 0)()(pWu
T
T
1 j
m
j
1
1
1
1
1
1
1
1
1
m
];)),()(1[( 1
0)(0)(1 1 ppWpWrv j
.
(4.1)
Differentiating the inter-temporal indirect utility function in (4.1) with
respect to jp :
jp
pWv
),( 0)(
)],([ 0)()(
1
)( pWu k
n
k
T
T
1 j
m
j
1
1
1
1
1
1
1
1
m
;
1
;
1
)(1
11
11 ),(j
j
W
pWv
k
j
W)(
;
11
j
k
p
pW
),( 0)(
T
T
1 j
m
j
1
1
1
1
1
1
1
1
m
;
1
;
1
)(1
11
11 ),(j
j
W
pWv
jp
Wj
;
11
, (4.2)
where jpWpWrWj
1
0)(0;
1 )),()(1(1
, k
t
j
W)(
;
11
kpr )1( and
jp
Wj
;
11
)1( r ),( 0)(pW
j
.
7324 David W. K.Yeung
Invoking the first order conditions in (3.8) the term inside the curly brackets
vanishes, condition (4.2) then becomes:
jp
pWv
),( 0)(
T
T
1 j
m
j
1
1
1
1
1
1
1
1
m
;
1
;
1
)(1
11
11 ),(j
j
W
pWv
)1( r ),( 0)(
pWj
.
(4.3)
The effect of a change in initial wealth on the maximized utility can be
obtained by differentiating ),( 0)(pWv
in (4.1) with respect to 0
W :
0
0)(),(
W
pWv
)],([ 0)()(
1
pWu k
n
k
T
T
1 j
m
j
1
1
1
1
1
1
1
1
m
;
1
;
1
)(1
11
11 ),(j
j
W
pWv
k
j
W)(
;
11
0
0)(),(
W
pWk
T
T
1 j
m
j
1
1
1
1
1
1
1
1
m
;
1
;
1
)(1
11
11 ),(j
j
W
pWv
)1( r .
(4.4)
Again, invoking the first order conditions in (3.8), the term inside the curly
brackets vanishes, condition (4.4) then becomes:
0
0)(),(
W
pWv
T
T
1 j
m
j
1
1
1
1
1
1
1
1
m
;
1
;
1
)(1
11
11 ),(j
j
W
pWv
)1( r .
(4.5)
Dividing the right-hand-side of equation (4.3) by the right-hand-side of
equation (4.5) and the left-hand-side of (4.3) by the left-hand-side of (4.5) yields:
jp
pWv
),( 0)(
0
0)(),(
W
pWv
),( 0)(pW
j
, for },,2,1{ nj . (4.6)
Random horizon stochastic dynamic Slutsky equation 7325
Condition (4.6) provides a random horizon stochastic dynamic version of Roy’s
identity involving a change in current prices under uncertain future preferences.
Then we consider deriving the random horizon stochastic dynamic Roy’s
identity under uncertain future preferences for a change in prices in current and
future periods.
Theorem 4.1. Random Horizon Stochastic Dynamic Roy’s Identity under
Uncertain Preferences
jp
pWv
),( 0)(
0
0)(),(
W
pWv
),( 0)(pW
j
, for },,2,1{ nj ; (4.7)
k
hp
pWv
),( 0)(
0
0)(),(
W
pWv
h
h
h
m
j
j
h
m
j
jm
j
j
11
2
1
1
2
2
2
1
1
1
2
2
1
1
1
1
2
1
1
mm
h
h
h
h
m
h
1
h
111
22
11
112
21
1
;
;)(),(
hhj
h
jj
hhj
h
jj
h
h
h
h
W
pWv
),( 112
21
1 ;)(pW h
hj
h
jj
h
h
k
h
)()1( hr
h
h
h
m
w
w
h
m
w
wm
w
w
11
2
1
1
2
2
2
1
1
1
2
2
1
1
1
1
2
1
1
mm
h
h
h
h
m
h
1
h
1
112
21
1
112
21
1
;
;)(),(
hhw
hww
hhw
hww
h
h
h
h
W
pWv
, (4.8)
for },,2,1{ T , },,2,1{ Th , },,2,1{ hnk and
},,2,1{ m ,
where 0
WW ,
;
1
11
j
W 1
1
0)(0 )],()[1(
j
pWpWr ,
12
21
1 ;
2
jj
W )],()[1(;
1
)(
11
;
1
111
11 pWpWr
jj
2
2
j ,
if preference is )( 1
)(1 1
xu
in period 1 ;
7326 David W. K.Yeung
213
32
21
1 ;
3
jjj
W )],()[1( 12
21
1112
21
1 ;
2
)(
22
;
2 pWpWrjjjj
3
3
j ,
if preference is )( 2
)(2 2
xu
in period 2 ;
112
21
1 ,;
TTj
Tjj
TW
)]()[1( 211
12
21
11211
12
21
1 ;
1
)(
11
;
1
TTj
T
jj
TTTj
T
jj
TTTT WpWr
Tj
T ;
if preference is )( 1
)(1 2
T
Txu in period 1T ;
TTj
Tjj
TW
1
11
22
11 ,;
1
)]()[1( 112
21
1112
21
1 ;)(;
TTj
Tjj
TTj
Tjj
TTTT WpWr
01
1
Tj
T .
(4.9)
Proof. See the Appendix. ■
Theorem 4.1 gives the random horizon stochastic dynamic Roy’s identity under
uncertain future preferences. Invoking (A.9) in the proof of Theorem 4.1 in the
Appendix, an alternative form of the random horizon stochastic dynamic Roy’s
identity can be expressed as:
k
hp
pWv
),( 0)(
T
T
h
h
h
h
m
j
j
h
m
j
jm
j
j
11
2
1
1
2
2
2
1
1
1
1
1
1
1
1
m
2
2
2
1
2
m
1
1 1
h
h
h
m
h
h
111
22
11
112
21
1
;
;)(),(
hhj
hjj
hhj
hjj
h
h
h
h
W
pWv
),( 112
21
1 ;)(pW h
hj
h
jj
h
h
k
h
,
(4.10)
for },,2,1{ T , },,2,1{ Th , },,2,1{ hnk and
},,2,1{ m .
5 Duality and Wealth Compensated Demand
In this section, we invoke the duality principle in consumer theory to construct
wealth compensated demand functions under an uncertain inter-temporal budget,
uncertain preferences and a random life-span. To do this, we consider the dual
Random horizon stochastic dynamic Slutsky equation 7327
problem of minimizing expenditure covered by the current wealth subject to
maintaining the level of utility achieved in the primal problem. Following the
analysis in Yeung (2014) we first examine the case when the consumer survives
in the last period and his preference is )()(
T
Txu T . Let *0
TT WW denote the
consumer’s wealth in period T . Since wealth equals income in this period, to
derive the compensated demand we follow the standard single period consumer
problem of
TTx
xpT
m i n
subject to achieving the level of utility )()(
T
Txu T
0
)(ˆ T
T
W
Tv = ),( 0)(
TT
TpWv T .
(5.1)
Setting the corresponding Lagrange function and performing the minimization
operation yields a set of first order conditions. With the implicit function
theorem holding for the first order conditions one can obtain the wealth (income)
compensated demand functions as
h
Tx ),ˆ(0
)(
)(
T
W
T
h
T pv T
T
T
, for },,2,1{ Tnh . (5.2)
Substituting (5.2) into (5.1) yields the wealth-expenditure function
),ˆ(0
)(
)(
T
W
TT pv T
T
T
),ˆ(0
)(
)(
T
W
TTT pvp T
T
T
0
TW .
Now we proceed to period 1T and let wealth in this period be *
1
0
1 TT WW
and preference be )( 1
)(1 1
T
Txu T . To obtain the wealth compensated demand
function in period 1T we consider the problem of minimizing expenditure
covered by current wealth in the period to bring about the expected inter-temporal
utility 0
1
1 )(1ˆ
T
T
W
Tv ),,( 1
0
1
)(1 1
TTT
TppWv T
from the primal problem. However,
wealth 0
1TW in period 1T does not only cover consumption expenditure
11 TT xp in the period 1T but also part of the consumption expenditure in
period T . To delineate expenditures attributed to wealth in period 1T we first
invoke the dynamical equation (2.3) and express 1TW as:
)()1( 1
111 TTTTT WrxpW
. (5.3)
Using the wealth-expenditure function ),ˆ(0
)(
)(
T
W
TT pv T
T
T
in period T and taking
expectation over the random variables T and T~ one can obtain a crucial
identity relating wealth to current and expected future expenditures attributable to
wealth from (5.3) as:
7328 David W. K.Yeung
11
0
1 TTT xpW ]),ˆ([)1())(1(
)(
)(
11
1 110
1 j
TT
xpWr
TT
m
T
m
j
j
T pvrj
TTTT
T
T
T
T
T
T
. (5.4)
Using (5.4) the consumer’s dual problem in period 1T can be formulated as
minimizing wealth expenditure
11 TT xp ]),ˆ([)1())(1(
)(
)(
11
1 110
1 j
TT
xpWr
TT
m
T
m
j
j
T pvrj
TTTT
T
T
T
T
T
T
(5.5)
with respect to 1Tx subject to the constraint
)( 1
1
T
T xu )],ˆ([))(1(
)(
)()(
11
1
110
1
T
xpWr
TT
T
T
m
T
m
j
j
TT
T
T pvuj
TTTT
T
TT
T
T
T
T
0
1
1 )(1ˆ
T
T
W
Tv .
(5.6)
Since ),ˆ())(1(
)(
)( 110
1
T
xpWr
TT pvj
TTTT
T
T
is a set of wealth compensated demands
that leads to the level of utility j
TTTT xpWr
Tv
))(1(
)(11
01ˆ , so
),ˆ([))(1(
)(
)()( 110
1
T
xpWr
TT
Tpvu
jTTTT
T
TT
equals
jTTTT
T
xpWr
Tv
))(1(
)(11
01ˆ . Invoking
jTTTT
T
xpWr
Tv
))(1(
)(11
01ˆ ];))(1[( 11
0
1
)(
T
j
TTTT
TpxpWrv T
the constraint (5.6)
can be expressed as:
)( 1
)(1 1
T
Txu T
Tm
j
j
TT
T
T
1
1
T
T
T
m
T
1
T ];))(1[( 11
0
1
)(
T
j
TTTT
TpxpWrv T
01
1 )(1ˆ
T
T
W
Tv .
Setting the Lagrange function and performing the relevant optimization
operation (similar to the analysis in Yeung (2014)) yields a set of first order
conditions. With the implicit function theorem holding, the wealth compensated
demand functions can be obtained as:
),,ˆ( 1)(1
)(
1
)(
1
01
1
11
TT
W
T
h
T
h
T ppvx T
T
TT
, for },,2,1{ 1 Tnh . (5.7)
Substituting the wealth compensated demand functions in (5.7) into (5.5)
yields the wealth-expenditure function in period 1T :
Random horizon stochastic dynamic Slutsky equation 7329
),,ˆ( 1)(1
)(
1
01
1
1
TT
W
TT ppv T
T
T
1 Tp ),,ˆ( 1)(1
)(
1
01
1
1
TT
W
TT ppv T
T
T
Tm
j
j
Tr1
1)1(
T
T
T
m
T
1
),ˆ(),,ˆ(
)(
)( 1
01
1110
1
T
ppvpW
TT pvj
TTTT
W
TTTT
T
T
. (5.8)
Now we proceed to period }1,,3,2{ TTk and let wealth be 0
kW *
kW
and preference be )()(
k
kxu k in the period. Again using (2.3) we can express
wealth in period k as )()1( 11
1
kkkkk WrxpW . Invoking the
wealth-expenditure functions in period 1k and taking expectations over the
random variables 1k and 1~
k , one can obtain the identity
kk xp
1
1
1
1
1
1)1(k
k
k
m
j
j
kr
1
1
1
1
1
k
k
k
m
k
110
1
1
1
))(1(
)(1
)(
1 ),ˆ(
kj
kkkk
k
k j
k
xpWr
kk pv
0
kW , (5.9)
where ),ˆ( 10
1
))(1(
)(1 pvj
kkkk
k
xpWr
k
is the short form for
),,,,ˆ( 21
))(1(
)(11
0
1 Tkk
xpWr
k pppvj
kkkk
k
.
The consumer’s wealth expenditure minimization problem can be expressed as:
kxmin kk xp
1
1
1
1
1
1)1(k
k
k
m
j
j
kr
1
1
1
1
1
k
k
k
m
k
11
10
1
1
1
))(1(
)(1
)(
1 ),ˆ(
kkj
kkkk
k
k j
k
xpWr
kk pv
(5.10)
subject to
)()(
k
kxu k
T
k
T
k
1
1
1
1
1
1
k
k
k
m
j
j
k
1
1
1
1
1
k
k
k
m
k
1k ];))(1[( 11
1
0)(1pxpWrv kk j
kkkk
k
0
)(ˆ k
k
W
kv , (5.11)
for }2,,2,1{ Tk and *0
kk WW .
Setting up the Lagrange function and deriving the first order conditions one
can obtain the wealth compensated demand functions (with the implicit function
7330 David W. K.Yeung
theorem holding) as ),ˆ(0
)(
)(pvx k
k
k W
k
h
k
h
k
, for }2,,2,1{ Tk ,
},,2,1{ knh and *0
kk WW .
Similarly, the wealth-expenditure function can be obtained as:
),ˆ(0
)(
)(pv k
k
k W
kk
kp ),ˆ(0
)(
)(pv k
k
k W
kk
1
1
1
1
1
1)1(k
k
k
m
j
j
kr
1
1
1
1
1
k
k
k
m
k
11
10
1
1
1
))(1(
)(1
)(
1 ),ˆ(
kkj
kkkk
k
k j
k
pWr
kk pv
. (5.12)
The wealth compensation demand functions and wealth-expenditure
functions derived in this section represent the dual results of the primal problem in
Section 3.
6 Random Horizon Stochastic Dynamic Slutsky Equations
In this section, we derive the Slutsky equations under a dynamic framework with
uncertainties in the consumer’s future income, future preferences and life-span.
Invoking the duality results in Section 3 and Section 5 we have
),ˆ(0
)(
)(pv
Wh
),( 0)(pW
h
, and ),ˆ(
0
)(
)(pv
W
0
W and
),( 0)(pWv
0
)(ˆ
W
v , for *0
WW and },,2,1{ T , },,2,1{ nh and
},,2,1{ m .
Substituting 0
W by ),ˆ(0
)(
)(pv
W
into the wealth-dependent ordinary
demand function yields the identity:
),ˆ(0
)(
)(pv
Wh
]),,ˆ([0
)(
)(ppv
Wh
(6.1)
for },,2,1{ nh and },,2,1{ m .
One can derive a theorem concerning the relationships between the price
effect of the demand of a commodity and the pure substation effect and the wealth
effect in a random horizon stochastic dynamic framework with uncertain future
preferences as follows.
Random horizon stochastic dynamic Slutsky equation 7331
Theorem 6.1. Random Horizon Stochastic Dynamic Slutsky Equation
under Preference Uncertainty
i
h
p
pW
),( 0)(
i
h
p
pW
),( 0)(0
0)(),(
W
pWh
),( 0 pW
i
,
ki
k
h
p
pW
),( 0)(
ki
k
h
p
pW
),( 0)(
0
0)(),(
W
pWh
m
j
j
k
m
j
jm
j
j k
11
2
1
1
2
2
2
1
1
1
2
2
2
1
1
1
1
2
1
1
mm
k
k
k
m
k
1
k
111
22
11
112
21
1
;
;)(),(
kkj
k
jj
kkj
k
jj
k
k
k
k
W
pWv
),( 112
21
1 ;pW k
kj
k
jj
k
k
i
k
)()1( kr
k
k
k
m
w
w
k
m
w
wm
w
w
11
2
1
1
2
2
2
1
1
1
2
2
2
1
1
1
1
2
1
1
mm
h
h
h
m
h
1
h
1
112
21
1
112
21
1
;
;)(),(
hhw
hww
hhw
hww
h
h
h
h
W
pWv
, (6.2)
for },,2,1{ T , },,2,1{ Tk , },,2,1{ kk ni , },,2,1{, nih
and
},,2,1{ m .
Proof. Differentiating the identity (6.1) with respect to ti
tp yields:
ti
t
Wh
p
ppv
]),,ˆ([0
)(
)(
ti
t
Wh
p
pv
),ˆ(0
)(
)(
),ˆ(
]),,ˆ([0
0
)(
)(
)(
)()(
pv
ppv
W
Wh
ti
t
W
p
pv
),ˆ(0
)(
)(
, (6.3)
for },,2,1{ tt ni , },,1,{ Tt and },,2,1{ m .
Invoking ),ˆ(0
)(
)(pv
W
0
W one can express (6.3) as:
ti
t
h
p
pW
),( 0)(
ti
t
Wh
p
pv
),ˆ(0
)(
)(
0
0
)(
)(),(
W
pWh
ti
t
W
p
pv
),ˆ(0
)(
)(
. (6.4)
7332 David W. K.Yeung
To derive the term ti
t
W
p
pv
),ˆ(0
)(
)(
in a more readily computable form we
first note that 0
)(ˆ
W
v ),( 0)(pWv
. To derive the effect on
),ˆ(0
)(
)(pv
W
brought about by a change in ti
tp , with 0
)(ˆ
W
v being held constant,
we totally differentiate 0
)(ˆ
W
v to obtain:
0
)(ˆ
W
vd 0
0
0)(),(
dWW
pWv
i
i
T n
j
dpp
pWv
),( 0)(
1
. (6.5)
With 0ˆ0
)(
W
vd and 0
idp for all },,2,1{ ni and },,1,{ T
except ti
tdp , equation (6.5) becomes
0 0
0
0)(),(
dWW
pWv k
t
t
i
ti
t
dpp
pWv
),( 0)(
,
which yields
ti
tdp
dW 0
t
k
i
tp
pWv
),( 0)(
0
0)(),(
W
pWv
0
)(
0
ˆ
Wi
tvp
Wt
. (6.6)
Invoking ),ˆ(0
)(
)(pv
W
0
W and using (6.6) one can readily obtain
ti
t
W
p
pv
),ˆ(0
)(
)(
0
)(
0
ˆ
Wi
tvp
Wt
t
k
i
tp
pWv
),( 0)(
0
0)(),(
W
pWv
. (6.7)
Substituting (6.7) into (6.4) and invoking the random horizon stochastic dynamic
Roy’s identity in Theorem 4.1, one obtains (6.2). Hence Theorem 6.1 follows.
■
The random horizon stochastic dynamic Slutsky equation under uncertain
preferences in (6.2) generalizes the classic Slutsky equation to a multi-period
framework with uncertainties in future income, the consumer’s life-span and
preferences. In particular, the effect of a price change on the demand of a
commodity can be decomposed into a pure substation effect and a wealth effect.
The left hand side of equation (6.2) represents how the demand for good h at
period changes in response to a change in price ti
tp , and the first term on the
right hand side of the equation gives the change in demand caused by a change in
price ti
tp holding utility fixed at 0
)(ˆ
W
v . The second term on the right hand side
Random horizon stochastic dynamic Slutsky equation 7333
of equation (6.2) is the product of the change in demand when wealth changes and
the required change in wealth brought about by a change in ti
tp with utility kept
fixed at 0
)(ˆ
W
v . Thus, the change in the demand of a good caused by a price
change can be decomposed into a pure substation effect and a wealth effect.
7 Concluding Remarks
This paper extends the conventional consumer analysis to a random horizon
stochastic dynamic framework in which there are uncertainties in the consumer’s
future incomes, future preferences and life-span. The extension incorporates
realistic and essential characteristics of the consumer into conventional consumer
theory. The paper derives two of the most crucial foundations in consumer
theory – Roy’s identity and Slutsky equation – in a random horizon stochastic
dynamic framework with uncertain future preferences. The analysis advances the
microeconomic study on optimal consumption decision to a random horizon
stochastic dynamic framework with uncertain consumer preferences. Further
research, development and propagations which explore further economic
implications of the results in this paper are in order.
8 Appendix: Proof of Theorem 4.1.
Invoking (3.7) we obtain the identity
),( 0)(pWv
)],([ 0)()(pWu
T
T
1 1
1
1
1
1
j
m
j
1
1
1
1
1
m
1 )],([;
1
)(
1
)(1 1111 pWu
j
T
T
2
2
2
2
1
1
1
1
2
1
1
m
j
jm
j
j
1
1
1
1
1
m
2
2
2
1
2
m
2
1
)],([ 1
22
1122 ;
2
)(
2
)(2pWu
jj
7334 David W. K.Yeung
T
T
h
h
h
h
m
j
j
h
m
j
jm
j
j
11
2
1
1
2
2
2
1
1
1
1
1
1
1
1
m
2
2
2
1
2
m
h
h
h
m
h
1
h
1 )],([ 112
21
1 ;)()(pWu h
hj
h
jj
hh
hh
h
T
T
h
1
1
1
1
2
2
2
1
1
1
1
1
1
2
1
1
h
h
h
m
j
j
h
m
j
jm
j
j
1
1
1
1
1
m
2
2
2
1
2
m
1
1
1
1
1
h
h
h
m
h
1
1
h
)(1 1 hh
v
),( 11
12
21
1 ;
1 pW hhj
h
jj
h
. (A.1)
Differentiating (A.1) with respect to k
hp yields:
k
hp
pWv
),( 0)(
)],([ 0)()(
1
pWu i
n
i
T
T
h
1
1
1
1
2
2
2
1
1
1
1
1
1
2
1
1
h
h
h
m
j
j
h
m
j
jm
j
j
1
1
1
1
1
m
2
2
2
1
2
m
1
1
1
1
1
h
h
h
m
h
1
1
h
h
hjh
jj
hhj
hjj
h
h
h
h
W
pWv
11
12
21
1
11
12
21
11
;
1
;
1
)(1),(
i
hh
hj
h
jj
W)(
;
11
11
22
11
k
h
i
p
pW
),( 0)(
h
1
m
j
jm
j
jm
j
j
11
2
1
1
2
2
2
1
1
1
1
1
1
1
1
m
2
2
2
1
2
m
m
1
n
i 1
T
T
1 )],([ 11
22
11
)(
;)()(pWu
jjj
i
Random horizon stochastic dynamic Slutsky equation 7335
T
T
h
1
1
1
1
2
2
2
1
1
1
1
1
1
2
1
1
h
h
h
m
j
j
h
m
j
jm
j
j
1
1
1
1
1
m
2
2
2
1
2
m
1
1
1
1
1
h
h
h
m
h
1
1
h
h
hjh
jj
hhj
hjj
h
h
h
h
W
pWv
11
12
21
1
11
12
21
11
;
1
;
1
)(1),(
i
hh
hjh
jj
W)(
;
11
11
22
11
k
h
h
i
p
pW hhj
h
jj
),( 11
12
21
1 ;
1
)(
T
T
h
1
1
1
1
2
2
2
1
1
1
1
1
1
2
1
1
h
h
h
m
j
j
h
m
j
jm
j
j
1
1
1
1
1
m
2
2
2
1
2
m
1
1
1
1
1
h
h
h
m
h
1
1
h
h
hjh
jj
hhj
hjj
h
h
h
h
W
pWv
11
12
21
1
11
12
21
11
;
1
;
1
)(1),(
k
h
h
p
W hhj
h
jj
11
12
21
1 ;
1 .
(A.2)
Using (4.8) we have
i
hh
hjh
jj
W)(
;
11
11
22
11
ih pr 1)1( and
k
h
h
p
W hhj
h
jj
11
12
21
1 ;
1 )1( r ),( 112
21
1 ,;)(pW h
hj
h
jj
h
h
k
h
. (A.3)
Substituting (A.3) into (A.2) yields
k
hp
pWv
),( 0)(
)],([ 0)()(
1
pWu i
n
i
T
T
h
1
1
1
1
2
2
2
1
1
1
1
1
1
2
1
1
h
h
h
m
j
j
h
m
j
jm
j
j
1
1
1
1
1
m
2
2
2
1
2
m
1
1
1
1
1
h
h
h
m
h
1
1
h
h
hjh
jj
hhj
hjj
h
h
h
h
W
pWv
11
12
21
1
11
12
21
11
;
1
;
1
)(1),(
ih pr 1)1(
7336 David W. K.Yeung
k
h
i
p
pW
),( 0)(
h
1
m
j
jm
j
jm
j
j
11
2
1
1
2
2
2
1
1
1
1
1
1
1
1
m
2
2
2
1
2
m
m
1
n
i 1
T
T
1 )],([ 11
22
11
)(
;)()(pWu
jjj
i
T
T
h
1
1
1
1
2
2
2
1
1
1
1
1
1
2
1
1
h
h
h
m
j
j
h
m
j
jm
j
j
1
1
1
1
1
m
2
2
2
1
2
m
1
1
1
1
1
h
h
h
m
h
1
1
h
h
hjh
jj
hhj
hjj
h
h
h
h
W
pWv
11
12
21
1
11
12
21
11
;
1
;
1
)(1),(
ih pr 1)1(
k
h
h
i
p
pW hhj
h
jj
),( 11
12
21
1 ;
1
)(
T
T
h
1
1
1
1
2
2
2
1
1
1
1
1
1
2
1
1
h
h
h
m
j
j
h
m
j
jm
j
j
1
1
1
1
1
m
2
2
2
1
2
m
1
1
1
1
1
h
h
h
m
h
1
1
h
h
hj
h
jj
hhj
h
jj
h
h
h
h
W
pWv
11
12
21
1
11
12
21
11
;
1
;
1
)(1),(
)1( r ),( 112
21
1 ;)(pW h
hj
h
jj
h
h
k
h
. (A.4)
Invoking (4.5) we obtain:
112
21
1
112
21
1
;
;)(),(
jjj
jjj
W
pWv
Random horizon stochastic dynamic Slutsky equation 7337
T
T
h
1
1
1
1
2
2
2
1
1
1
1
1
1
2
1
1
h
h
h
m
j
j
h
m
j
jm
j
j
1
1
1
1
1
m
2
2
2
1
2
m
1
1
1
1
1
h
h
h
m
h
1
1
h
h
hjh
jj
hhj
hjj
h
h
h
h
W
pWv
11
12
21
1
11
12
21
11
;
1
;
1
)(1),(
1)1( hr . (A.5)
Using (A.5) the terms inside the square brackets in (A.4) can be written as
T
T
1 )],([ 11
22
11
)(
;)()(pWu
jjj
i
T
T
h
1 1
1
1
1
1
j
m
j
1
1
1
1
1
m
1
11
12
21
1
11
12
21
11
;
1
;
1
)(1),(
jjj
jjj
W
pWv)1( r
jp
.
(A.6)
Invoking the first order conditions in (3.8) the term inside the square brackets
in (A.6) vanishes and therefore (A.4) becomes:
k
hp
pWv ),( 0)(
T
T
h
1
1
1
1
2
2
2
1
1
1
1
1
1
2
1
1
h
h
h
m
j
j
h
m
j
jm
j
j
1
1
1
1
1
m
2
2
2
1
2
m
1
1
1
1
1
h
h
h
m
h
1
1
h
h
hjh
jj
hhj
hjj
h
h
h
h
W
pWv
11
12
21
1
11
12
21
11
;
1
;
1
)(1),(
)1( r ),( 112
21
1 ;)(pW h
hj
h
jj
h
h
k
h
. (A.7)
Using (A.5), one has
1
1
1
1
1
h
h
h
m
j
j
h
1
1
1
1
1
h
h
h
m
h
1hh
hjh
jj
hhj
hjj
h
h
h
h
W
pWv
11
12
21
1
11
12
21
11
;
1
;
1
)(1),(
7338 David W. K.Yeung
T
h
T
h
1
112
21
1
112
21
1
;
;)(),(
hhj
hjj
hhj
hjj
h
h
h
h
W
pWv
1)1( r . (A.8)
Substituting (A.8) into (A.7) yields
k
hp
pWv
),( 0)(
T
T
h
h
h
h
m
j
j
h
m
j
jm
j
j
11
2
1
1
2
2
2
1
1
1
1
1
1
1
1
m
2
2
2
1
2
m
1
1 1
h
h
h
m
h
h
111
22
11
112
21
1
;
;)(),(
hhj
h
jj
hhj
h
jj
h
h
h
h
W
pWv
),( 112
21
1 ;)(pW h
hj
h
jj
h
h
k
h
.
(A.9)
Invoking (A.5) one obtains
0
0)(),(
W
pWv
T
T
h
h
h
h
m
j
j
h
m
j
jm
j
j
11
2
1
1
2
2
2
1
1
1
1
1
1
1
1
m
2
2
2
1
2
m
1
1 1
h
h
h
m
h
h
111
22
11
112
21
1
;
;)(),(
hhj
hjj
hhj
hjj
h
h
h
h
W
pWv
hr)1( .
(A.10)
Dividing (A.9) by (A.10) yields another form of the random horizon Roy’s
identity as:
k
hp
pWv
),( 0)(
0
0)(),(
W
pWv
h
h
h
m
j
j
h
m
j
jm
j
j
11
2
1
1
2
2
2
1
1
1
h
h
h
h
m
h
mm
11
2
1
1
2
2
1
1
1
h
1
11
22
11
112
21
1
;
;)(),(
hhj
hjj
hhj
hjj
h
h
h
h
W
pWv
Random horizon stochastic dynamic Slutsky equation 7339
),( 112
21
1 ;)(pW h
hj
h
jj
h
h
k
h
)()1( hr
h
h
h
m
w
w
h
m
w
wm
w
w
11
2
1
1
2
2
2
1
1
1
2
2
1
1
1
1
2
1
1
mm
h
h
h
h
m
h
1
h
1
hw
h
ww
hhw
h
ww
h
h
h
h
W
pWv
22
11
112
21
1 ),(;)(
, (A.11)
for },,2,1{ T , },,2,1{ Th , },,2,1{ hnk and
},,2,1{ m .
Hence Theorem 4.1 follows. Q.E.D.
Acknowledgements. Financial support by HKSYU is gratefully acknowledged.
References
[1] R.G.D. Allen, Professor Slutsky’s Theory of Consumers’ Choice. Review of
Economic Studies, 3(1936), 120-129.
[2] R.G.D. Allen, the Work of Eugen Slutsky. Econometrica, 18(1950),
209-216.
[3] R. Bellman, Dynamic Programming. Princeton, Princeton University Press,
1957.
[4] M.T. Cheung and D.W.K. Yeung, Microeconomic Analytics, New York,
Prentice Hall, 1995.
[5] P.C. Dooley, Slutsky’s Equation Is Pareto’s Solution. History of Political
Economy, 15(1983), 513-517.
[6] T.W. Epps, Wealth Effects and Slutsky Equations for Assets. Econometrica,
43(1975), 301-303.
[7] J.R. Hicks and R.G.D. Allen, A Reconsideration of The Theory of Value.
Parts 1-2. Economica, New Series, 1(1934), 52-76 & 196-219. (Reprinted In
Hicks 1981, pp. 114-32).
7340 David W. K.Yeung
[8] R. Roy, La Distribution Du Revenu Entre Les Divers Biens. Econometrica,
15(1947), 205-225.
[9] H. Schultz, Interrelations of Demand, Price, and Income. Joumal of Political
Economy, 43(1935), 433-81.
[10] E.E. Slutsky, Sulla Teoria Del Bilancio Del Consumatore. Giornale Degli
Economisti, 51(1915), 1–26. Translated as “On the Theory of the Budegt of
the Consumer” in G. Stigler and K. Boulding (eds.), Reading in Price Theory,
Published for the Association by Richard D. Irwin, Inc., Homewood, IL,
1952.
[11] H. Varian, Microeconomic Analysis, 3rd ed., W.W. Norton, New York, 1992.
[12] D.W.K. Yeung, Optimal Consumption under an Uncertain Inter-temporal
Budget: Stochastic Dynamic Slutsky Equations. Vietsnik St Petersburg
University: Mathematics, 10(2013), 121-141.
[13] D.W.K. Yeung, Optimal Consumption under Uncertainties: Random Horizon
Stochastic Dynamic Roy’s Identity and Slutsky Equation, forthcoming in
Applied Mathematics, (2014).
[14] D.W.K. Yeung and L. A. Petrosyan, Subgame Consistent Cooperative
Solution of Dynamic Games with Random Horizon. Journal of Optimization
Theory and Applications, 150(2011), 78-97.
[15] D.W.K. Yeung and L. A. Petrosyan, Subgame-consistent cooperative
solutions in randomly furcating stochastic dynamic games. Mathematical and
Computer Modelling, 57(2013), 976–991.
[16] D.W.K. Yeung and L. A. Petrosyan, Subgame-consistent Cooperative
Solutions for Randomly Furcating Stochastic Dynamic Games with
Uncertain Horizon, forthcoming in International Game Theory Review,
(2014).
Received: August 23, 2014