You are on page 1of 10

Journal of Financial Risk Management, 2014, 3, 19-28

Published Online June 2014 in SciRes. http://www.scirp.org/journal/jfrm
http://dx.doi.org/10.4236/jfrm.2014.32003

Continuous-Time Mean-Variance Portfolio
Selection with Inflation in an Incomplete
Market
Yingying Xu1, Zhuwu Wu1,2
1

School of Science, China University of Mining and Technology, Xuzhou, China
School of Management, China University of Mining and Technology, Xuzhou, China
Email: xuyy23@hotmail.com, wuzhuwu@cumt.edu.cn

2

Received 14 April 2014; revised 10 May 2014; accepted 3 June 2014
Copyright © 2014 by authors and Scientific Research Publishing Inc.
This work is licensed under the Creative Commons Attribution International License (CC BY).
http://creativecommons.org/licenses/by/4.0/

Abstract
This paper concerns a continuous-time portfolio selection problem with inflation in an incomplete
market. By using the approach of more general stochastic linear quadratic control technique (SLQ),
we obtain the optimal strategy and efficient frontier to this problem. Furthermore, a numerical
example is also provided.

Keywords
Portfolio Selection, Efficient Frontier, Optimal Strategy, Stochastic Linear-Quadratic Control

1. Introduction
Portfolio selection problem is a key topic in the modern finance. The seminal work of Markowitz (1952, 1959)
addressed the issue of allocation of wealth in order to obtain the optimal return-risk trade-off. Since then, the
mean-variance model has been extended in many aspects. Merton (1969, 1971) introduced a continuous-time
model for maximizing the expected utility from investor’s consumption and terminal wealth. Zhou & Li (2000)
investigated a continuous-time mean-variance portfolio problem and obtained the optimal strategy and efficient
frontier by using the stochastic LQ technique, which opened up possible approach to solve the problem for more
constraints. Following Zhou & Li (2000), many scholars extended this model to the more complicated market
situations, such as liability, bankruptcy prohibition and incomplete market. See more details in Bielecki, Jin,
Pliska, & Zhou (2005), Xie, Li, & Wang (2008) and Ji (2010).
In a real world, investors must deal realistically with the problems of inflation with the growth of economy
when adopting a long-term but finite horizon investment strategy. Therefore, the consideration of inflation risk
How to cite this paper: Xu, Y. Y., & Wu, Z. W. (2014). Continuous-Time Mean-Variance Portfolio Selection with Inflation in
an Incomplete Market. Journal of Financial Risk Management, 3, 19-28. http://dx.doi.org/10.4236/jfrm.2014.32003

It is reasonable that the other uncertainties can be represented by one Brownian motion. T ] . .T ] . Stm satisfy the following stochastic differential equations: ( ) dSti = Sti bti dt + σ ti dWt . (4) The nominal price of real consumption goods in the economy at time t is denoted by Ψ t . In general. Y. Section 3 provides a closed-form solution of our model by using the more general stochastic LQ approach. On the other hand. The paper proceeds as follows. σ ) (σ ) (σ= 1 t 2 t m t ij t m×m .  Ψ (5)  t Ψ = 1. Section 4 presents a numerical example. However.  . t ∈ [ 0. σ ti := (σ ti1 . which represents the random factors ( ) 20 . concluding remarks and suggestions for future work are given in Section 5. our focus will be on two cases. The existing literature on this topic is not much as can be seen Brennan & Xia (2002) and Bensoussan. Keppo. bti is the appreciation rate of the ith ( i = 1. σ tim ) is the volatility associated with the ith risky asset. Xu. the model is formulated. The main goal of this paper is to investigate a continuous-time portfolio selection problemunder inflation in an incomplete market. which follows a diffusion process:  dΨ t = ρt dt + ξt dWt =: ρt dt + ξt1dWt1 +  + ξtm +1dWt m +1 . On the one hand. (3) where the superscript “ ′ ” represents the transpose of a vector or a matrix. we employ a stochastic linear quadratic (LQ) technique introduced by Zhou & Li (2000) to solve this problem. .Y. ∀t ∈ [ 0. we investigate the incomplete market with inflation. The remaining m assets are risky and their nominal price processes St1 . The price processes of risky assets are driven by an m-dimensional Brownian motion. Z. especially for long-term investors. bt2 . However. we assume the non-degeneracy condition of σ ( t ) σ ( t )′ ≥ ε I . by using the more general stochastic LQ control technique in Yong & Zhou (1999). but also with other uncertainties. which can be described by m + 1 Brownian motion. 2. It is clear that this model is more suitable and practical in most of the real-world situations. = (1)  0  S0= s0 > 0. which is our (m + 1)-th Brownian motion. to our knowledge. . In Section 2. It should be pointed out that the introduction of inflation is by no means routine and does give rise to difficulties which are not encountered in Zhou & Li (2000). where rt > 0 is the nominal interest rate of the risk-free asset. ε > 0. the covariance matrix of risky assets is as follows: ( ) = σ (t ) : ′ . W. We also assume that the inflation factors affected by the market are random. One of the assets is the risk-free whose nominal price process St0 is subject to the following ordinary differential equation: dSt0 rt St0 dt . σ ti 2 . . in which there are m risky assets and one risk-free asset. Finally. Wu in a portfolio selection model will make it more practical. the changes in the nominal price index are not just correlated with the risky assets’ nominal prices. we can also obtain the optimal strategy and efficient frontier in closed forms. Thus.  i  S0= si ∈ R (2) ′ where Wt := Wt1 . As widely adopted in the literature. The original idea can be seen in Brennan & Xia (2002). T ] . btm ) . m ) risky asset. Therefore. Wt is a (m + 1)-dimensional Brownian motion. & Sethi (2009). T ] . which represents the random factors that affect risky assets’ nominal prices. σ . let bt := ( bt1 . Wt m is a m-dimensional standard Brownian motion.  0 1 m +1 ′ where Wt := Wt . the research on mean-variance portfolio selection under inflation is limited. . . t ∈ [ 0. Problem Formulation We consider a market in which m + 1 assets are traded continuously within the time horizon [ 0.

in which the foregoing m Brownian motions are the same ones that drive the risky assets’ nominal price. . In general. F . We assume that all the coefficient functions are continuous bounded deterministic functions on [ 0.. Wt m +1 . Wsm +1 . P. 0 ′ . We call the process= π : {π t : t ∈ [ 0. . and the (m + 1)-th Brownian motion Wt m +1 represents other randomness. T ]} a portfolio or a strategy of the investor. t ≥ 0} and Let ( Ω. and let π t := π t1 . we obtain ( ) ( )    σt  Bt′ − σ t ξt′ − ξt ht  dWt − ξtm +1ht dWt m +1 dht =  rt − ρt + ξtξt′ ht + π t′  dt +  π t′ Ψt    Ψt     ′t − σ t ξt′  σt   B − ξt ht  dWt  dt +  π t′  =  rt − ρt + ξtξt′ ht + π t′ Ψt    Ψt   h0 = x  1 where π t = π t . We denote by C [ 0. .T ] . The real value of the investor’s wealth is given by X t Ψ t . R n×k the class of R n×k -valued continuous bounded deterministic functions on [ 0. we assume the shares of the (m + 1)-th risky asset held by the investor remains 0. R m the class of all R m -valued.T ] under P with norm = ξ L2 : F ( T E ∫ ξt dt 0 2 ) 1 2 < ∞. btm − rt . Applying Itô’s formula to ht . π tm . 2. m We denote by X t the nominal wealth of the investor at time t ∈ [ 0. Z. Bt = ( bt1 − rt . we give π t . Therefore. ξtm +1 ) is the volatility of the price index. Wu that affect the price index. T ] . T ] . W. and the (m + 1)-th risky asset is a fictitious risky asset. We assume that the trading of shares takes place continuously in a self-financing fashion and there are no transaction costs or taxes. Then Xt = m ∑ Nti Sti . We describe randomness of the price index with Wt1 . Moreover. we can 21 . . 0 ≤ t ≤ T. T ]. Remark 1. The first m risky assets are the same ones we assumed before. and by L2F [ 0.0 ≤ s ≤ t . Let = ht : X t Ψ t . Bt . progressively measurable and ( ( ) ) square integral random variables on [ 0.Y. Y. where= { } = Ft σ Ws . . { Ft }t ≥0 ) be a complete filtered probability space. ρt is the expected rate of inflation. F { Ft . Then we have ( ) dX t = rt X t dt + Bt π t dt + π t′σ t dWt   X 0= x > 0 (7) where Bt = ( bt1 − rt . . the driving factors of inflation include but do not equal to the ones of the risky assets’ nominal prices. 0 ) and ( (8) ) = σt  σ t11 σ t12  21 σ t22  σt ij   = σ  t m +1 × m +1 ( ) ( )  m1 m2 σ σ  t t  0 0  ( )  σ t1m  σ t2 m    σ tmm 0  0  0 . Moreover.T ] . we can also think that: it is assumed that there are m + 1 risky assets in the market. Xu. We also assume that short-selling is allowable. σ t . In fact. R ) . ∀ξt ∈ L2F ([0. the real value of any asset in the economy at time t is determined by deflating by the price index Ψ t . btm − rt ) is the risk premium. . The= ith ( i 1.1. T ] . π tm . With the consideration of the inflation. m ) asset held by the investor at time ′ t..1. m ) are independent. we assume that Wt m +1 and Wt j ( j = 1. m + 1) risky asset’s volatility is given by the ith rank of σ t . m ) at time t. and ξt := (ξt1 . 2. and their nominal prices are driving by m + 1 Brownian motions.  0 1  Remark 2. (6) i =0 Let π ti = N ti Sti be the total nominal market value of the ith ( i = 0. In order to facilitate the following mathematical treatment. . Suppose the investor decides to hold N ti shares of ith asset ( i = 0.

W. Wu get π t by deriving π t . we will introduce a stochastic LQ auxiliary control problem and derive its optimal feedback control. − ∞ < λ < +∞ . Y. Solution to the Problem In this section. λ )} (13) Recall Theorem 3. Theorem 1.Y. Firstly. For any µ > 0 .λ ) = {π π is an optimal control of A ( µ . (11) 3. The admissible strategy set under inflation with initial wealth x is defined as U= ( x) : {π π ∈ L ([0. λ ) . λ ) is equivalent to minimizing 1  J (π . we will apply the more general stochastic linear quadratic (LQ) control technique in Yong & Zhou (1999) to our model. λ ) ( min E µ hT2 − λ hT π ∈U ( x ) ) (12) where µ > 0. and at the same time to minimize the variance of the terminal real wealth. Var hT = E ( hT − EhT ) = EhT2 − ( EhT ) . Z. Then Equation (8) becomes the following stochastic differential equation: t ( ) ) (   σ  Vt dt +  π t′ t − ξt Yt − γξt  dWt dYt = rt − ρt + ξtξt′ Yt dt + rt − ρt + ξtξt′ γ dt + π t′ Ψt   Ψt   = = − : Y y x γ  0 (14) where V= Bt′ − σ t ξt′ . R ) . Let γ = λ ( 2 µ ) . γ ) := E  µYT2  2  22 (15) . the auxiliary problem A ( µ . and the objective function of the auxiliary problem A ( µ . VarhT . 2 2 This is the mean-variance model which can be expressed by the bi-objective optimization problem: P min ( − EhT . t m +1 2 F t t The objective of the investor is to maximize the expected terminal real wealth. π ) satisfies Equation (8)} . Auxiliary Problem Similar to Zhou & Li (2000). Xu. Eventually the optimal portfolio strategy and the efficient frontier for the original mean-variance portfolio optimization problem under inflation are obtained in closed form. π ∈U ( x ) (10) where the parameter µ > 0 represents the weight imposed by the investor on the objective VarhT . 3. Y= ht − γ .1 in Zhou & Li (2000) which shows the relationship between problems P ( µ ) and A( µ.VarhT ) . λ ) becomes E ( µYT2 − µγ 2 ) . then π ∈ ∏ A( µ . (9) π ∈U ( x ) It is known from Li & Ng (2000) that Equation (9) is equivalent to the following single objective optimization problem: P(µ ) min ( − EhT + µVarhT ) . t Hence. Define ∏ A( µ . ( h .1.λ ) .λ ) with λ ∗ = 1 + 2 µ EhT∗ . EhT . T ]. where ht∗ is the wealth process corres- ponding to the strategy π ∗ . ∗ ∗ Moreover. if π ∈ ∏ P( µ ) . one has ∏ P( µ ) ⊆  −∞< λ <+∞ ∏ A( µ . we introduce an auxiliary problem as follows: A( µ. Define ∏ P( µ ) := {π π is an optimal strategy of P ( µ )} .

D )′ . Substituting Equation (18) into Equation (14). = Pt : = . in an essential way. Ψt ( D . T ] . t m +1 2 F t t Thus the auxiliary problem A ( µ . W. λ ) is equivalent to the following stochastic LQ control problem: A (γ ) min J (π . T .. γ ) . (18) Moreover. respectively. π ) satisfies Equation (14)} . R + ) and gt ∈ C ([ 0. (17) where dPt dg t . ∀t ∈ 0. −1 Then Problem A ( γ ) is solvable with the optimal control being in a state feedback form.Y. Dt = Ψt Vt . Y. 2   (19) where M = ∑ i . T ] . T ]. . we have  dYt = ( At − M tϕt ) Yt + γ At − M tψ t  dt + (ξt − ϕt′Dt′ ) Yt + γξt −ψ t′Dt′  dWt   Y0 = y. −1 ′ ′ ′ ( Dt′PD t t ) ( M t g t + γ Dt Pt ξ t ) . π ∈Λ ( y ) 3. M t = Dt j = ( ) 1 σ t1 j . Solution to the Auxiliary Problem A solution of the stochastic LQ control problem A ( γ ) will involve. (Y . [ ]  t t t  (16) along with the following adjoint ordinary differential equation:    −1 −1 ′ 0  g t +  At − M t ( Dt′Dt ) ( M t′ + Dt′ξt′ )  gt +  ξt′ − Dt ( Dt′Dt ) ( M t′ + Dt′ξt′)  γξt′ + γ At  Pt =      gT = 0. Proof: We first prove that the control given by Equation (18) is an admissible control. π * (Yt ) = −ϕtYt −ψ t . Xu. T ] . Wu Furthermore. Z. σ tm +1 j .2. m +1 t 1 t Theorem 2. j mij2 for any matrix or vector M = ( mij ) and y= x − γ . the admissible strategy set U ( x ) can be written as Λ ( y ) := {π π ∈ L ([0. R ) . R ) be the solution of Equations (16) and (17). 23 (20) . such that = ϕt : = ψt : ( Dt′Dt ) ( M t′ + Dt′ξt′) . t ∈ [ 0. the optimal cost value is = J* 1 1 T  2 2γ At gt + γ Pt ξt ξt′ − ( Dt′Pt Dt ) 2 ψ t ∫ 0 2   2  1  2  dt + P0 y + g 0 y . the following Riccati equation:   −1  ′ 0  Pt +  2 At + ξt ξt′ − ( M t′ + Dt′ξt′ ) ( Dt′Dt ) ( M t′ + Dt′ξt′ )  Pt =     PT = µ  D ′P D > 0. g t : dt dt At =rt − ρt + ξt ξt′ . Let Pt ∈ C ([ 0.

T ] . Therefore. λ ) is also given by Theorem 2: π * ( ht ) = −ϕt ( ht − γ ) −ψ t = − ( Dt′Dt )   −1 ( M ′ + D′ξ ′ ) ( h − γ ) − ( D′D ) ( M ′H + γ D′ξ ′ ) . we have shown that π ∗ (Yt ) ∈ Λ ( y ) . and integrating them from 0 to T. T ] . Solution to the Original Problem Let ht* be the wealth process under the optimal feedback control π * of the auxiliary problem A ( µ . and the appreciation coefficient and the diffusion coefficients are bounded continuous within t. The proof is completed. γ ) − 1  1 1 T P0Y02 − g 0Y0 ≥ ∫0 2γ At gt + γ 2 Pt ξt ξt′ − ( Dt′Pt Dt ) 2 ψ t 2 2   2    dt . taking expectations. let Yt 2 be the state variable associated with the control vector π t . Substituting π * ( ht∗ ) = −ϕt ( ht∗ − γ ) −ψ t into Equation (8) yields  *   ′ * * * dht= ( At − M tϕt ) ht + ( γ M tϕt − M tψ t )  dt +  −ϕt ht − γ −ψ t Dt′ + ht ξt  dWt     h∗ 0 = x ( )  0 ( ( 24 ) ) (25) . λ ) .  Noting that the third constraint in Equation (16) is satisfied automatically since the assumption σ tσ t′ ≥ ε I m . −1 t t t t t t t t t t t ∈ [ 0. gt are continuous. W. For any π t ∈ Λ ( y ) .T ] 2 ). one has −1 T − ∫  As +ξ sξ s′ −ξ s Ds ( Ds′ Ds ) ( M s′ + Ds′ ξ s′ )  ds    Ht = γ 1 − e t  ⋅   (23) Since the equivalence of problem A ( γ ) and A ( µ . It shows that the feedback control given by Equation (18) is an optimal control and the optimal cost function can be obtained by Equation (19). the optimal feedback control of the auxiliary problem A ( µ . Z. we get 1 1 1 µYT2 − P0Y02 − g 0Y0 = J (π . 2     (21) Because 1 T E (π t + ϕtYt +ψ t )′ ( Dt′Pt Dt )(π t + ϕt Yt +ψ t ) dt ≥ 0 . we have π * ( ht ) = − ( Dt′Dt )   −1 ( M ′ + D′ξ ′ ) h + γ ( D′D ) t t t t t −1 t −1 T − ∫t  As +ξ sξ s′ −ξ s Ds ( Ds′ Ds ) ( M s′ + Ds′ ξ s′ )  ds      + Dt′ξt′  − Dt′ξt′  . Obviously. Xu. T ] .       (22) Let H t = gt Pt . λ ) . the solution of Equation (16) can be expressed by the following: 2 1     T  Pt= λ exp − ∫t  2 As + ξ sξ s ′ − ( Ds′ Ds ) 2 ϕ s  ds  .   and the equality holds if and only if π (Yt ) ≡ −ϕtYt −ψ t . (24) Substituting Equation (23) into Equation (24). Then noting Equation (16) and Equation (17).3.Y. Y. 2 ∫0 We obtain J (π . Wu Noting that Pt .  M t′e    3. γ ) − P0Y02 − g 0Y0 2 2 2 2 1   1 T  = E ∫0 (π + ϕt Yt +ψ t )′ ( Dt′Pt Dt )(π t + ϕt Yt +ψ t ) − ( Dt′Pt Dt ) 2 ψ t + 2γ At gt + γ 2 Pt ξt ξt′  dt. add them together. ∀t ∈ [ 0. By applying Itô formula to PY t t 2 and g t Yt . Next. we deduce that Equation (20) admits a unique strong solution Yt ∗ which yields ( E sup Yt* ≤ KT 1 + y t∈[ 0. where KT > 0 is a constant associated with the terminal time. we prove that π * (Yt ) is an optimal feedback control of state variable Yt . Hence. t ∈ [ 0.

which leads respectively to dEht* = ( At − M tϕt ) Eht* + ( γ M tϕt − M tψ t )  dt  ∗  Eh0 = x (27) dEh∗2 =  2 A − M ( D′D )−1 M ′ − 2 M ( D′D )−1 D′ξ ′ − ξ D ( D′D )−1 D′ξ ′ + ξ ξ ′  Eh∗2 dt t t t t t t t t t t t t t t t t t t   t  t   g  2  g  −1 −1 −1  t  M t ( Dt′Dt ) M t′ + γ 2 M t ( Dt′Dt ) M t′ − 2γ M t t ( Dt′Dt ) M t′  dt +  Pt  Pt     ∗2 2  Eh0 = x  (28) and The solution of Equation (27) is ∗ Eh = t t t x0 ∫0t ( As − M sϕs )ds ( A − M ϕ ) dz e + ∫0 ( γ M sϕ s − M sψ s ) e ∫s z z z ds. T ] . Wu Applying Itô’s formula to ht∗2 yields dh∗2 =  2 ( A − M ϕ ) + ϕ ′D′D ϕ − 2ξ D ϕ + ξ ξ ′  h∗2 dt +  2 ( γ M ϕ − M ψ ) − 2γϕ ′D′D ϕ t t t t t t t t t t t t t t t t t t t t   t   t  + 2ψ t′Dt′Dtϕt + 2 ( γξt Dtϕt − ξt Dtψ t )  ht*dt + γ 2ϕt′Dt′Dtϕt − 2γψ t′Dt′Dtϕt +ψ t′Dt′Dtψ t dt     ′  + 2ht*  −ϕt ht* − γ −ψ t Dt′ + ht*ξt  dWt     ∗2 2 h0 = x ( ( ( ) ) ) (26) Taking expectation on both side of Equations (25) and (26). q0 This leads to EhT∗= α + βγ . (30) t ∈ [ 0. we have −1 −1 T ∫  2 At − M t ( Dt′Dt ) M t′ − 2 M t ( Dt′Dt ) Dt′ξt′  dt = EhT∗2 x 2 e 0  −1 −1 T ∫  2 At − M t ( Dt′ Dt ) M t′ − 2 M t ( Dt′ Dt ) Dt′ξt′  dt = x 2 e 0  + γ 2 ∫0 M t ( Dt′Dt ) M t′e T + γ 2 1 − e  −1 −1 T − ∫0 M t ( Dt′ Dt ) M t′dt −1 − ∫t M s ( Ds′ Ds ) M s′ ds T dt   . β := 1 − e −1 T − ∫0 M t ( Dt′ Dt ) M t′dt .Y.  Then we get EhT∗2= η + βγ 2 . 25 (31) . Y. t ∈ [ 0. −1 −1 T ∫  2 At − M t ( D′D ) M t′ − 2 M t ( D′D ) Dt′ξt′  dt where η := x 2 e 0  . Substituting Equation (23) into Equation (30). T ] . t ∈ [ 0. T ] . Xu. W. (29) where α := xe ∫0 ( T At − M tϕt )dt . by solving Equation (28) we have −1 −1 T ∫  2 At − M t ( Dt′Dt ) M t′ − 2 M t ( Dt′Dt ) Dt′ξt′  dt EhT∗2 = x 2 e 0  2  T  g  g −1 −1 −1 + ∫0  t  M t ( Dt′Dt ) M t′ + γ 2 M t ( Dt′Dt ) M t′ − 2γ M t t ( Dt′Dt ) M t′  Pt  Pt   −1 −1 T ∫  2 As − M s ( Ds′ Ds ) M s′ − 2 M s ( Ds′ Ds ) Ds′ ξ s′  ds ×e t dt . Similarly. Z.

0. Numerical Example In this section. And also suppose that the covariance matrix σ t ( )  0.1772 0. According to EhT*= α + βγ .1432   0. then Equation (34) would reduce to − ∫0 Bt (σ t′σ t ) T = Varh * T e −1 Bt′dt −1 T − ∫0 Bt (σ t′σ t ) Bt′dt ( T EhT* − xe ∫0 rt dt 1− e Obviously.0823. If we let = ρt 0.1127 0.0.2656 0. β .0310 .1621.1422. if any optimal solution of problem P ( µ ) exists. 1− β Thus. the efficient frontier of our model is obtained by the following. the appreciation rate of assets b = ( 0. the volatility of the price index ξ = ( 0.0. W.1127  0.η into Equation (32). the time horizon T = 2. EhT*= α + βγ * with γ * = λ * 2 µ . it can be obtained by the solution π * of the auxiliary problem A ( µ .= ξ 0 .0956 0. Z. By substituting = γ∗ ( Eh −α ∗ T ) 2 (32) β and α . we finally obtain the efficient frontier as follows: −1 Varh = * T e −1 T − ∫0 M t ( Dt′ Dt ) M t′dt 1− e −1 − ∫0 M t ( Dt′ Dt ) M t′dt T T ( A − M ϕ ) dt ) 2 xe ∫0 t t t −1 −1  T ∫0  2 At − M t ( D ′D ) M t′ − 2 M t ( D ′D ) Dt′Ct  dt 2  +x e − T −1 − M ( D ′ D ) M ′dt  e ∫0 t t t t  −1 ( M ′g + γ D′Pξ ′) t t and Equation (23) into Equation t t t V ′ (σ ′σ )−1V dt −1 T  * ∫0 ( At +Vt′ (σ t′σ t ) σ t′ξt′ )dt   EhT − xe  . 26 .0350 .2372  4×4 After some transformations and calculations by using the above parameters.1396.4426 . (34)   −1 T  * e ∫0 t t t t ∫0 ( At − M t ( D′D ) D′ξ ′ )dt  −1  EhT − xe  = T   1 − e − ∫0 Vt′ (σ t′σ t ) Vt dt − 2 T (33) 2 Remark 3. we have 2 = VarhT* = 1− β  * α  α2  EhT −  +η − 1− β  1− β β  e −1 T − ∫0 M t ( Dt′ Dt ) M t′dt 1− e − ∫0 M t ( Dt′ Dt ) T −1 M t′dt T ∫0 ( At − M t ϕt ) dt   EhT* − xeT −1 − M ( D ′ D ) M ′dt  e ∫0 t t t t  ( ) ( 2 = ϕt ( Dt′Dt ) = M t′ + Dt′ξt′ .0.1326 ) i. 0. and the market is complete.0236 Eh2* − 1. j is as follows: the initial wealth x0 = 1 .1831   0.2123.1856 0. the result of Zhou & Li (2000) is a special case in our paper.1562. + 2 µ (1 − β ) 1 − β * γ γ= = Correspondingly. λ * ) with λ * = 1 + 2µ EhT* . ).0. Let’s assign the following parameters which are needed in our model: the risk-free asset rt = 0.1432 0.0956 0.1013. 2 (35) 4. The above two equations yield λ* = 1 + 2 µα .1876 ) .Y. Y.1362   0.3121   0. Wu Based on the Theorem 1. the variance of the terminal wealth is ( ) VarhT∗= EhT∗2 − EhT∗ = β (1 − β ) γ ∗2 − 2αβγ ∗ + η − α 2 .1831 0. the expected value of inflation ρt = 0.0. Xu. Suppose that the market has four assets and a risk-free asset.1362 0. π * ( ht ) = with 1 α . we discuss a numerical example. the optimal feedback control of the problem P ( µ ) can be expressed by −ϕt ( ht − γ ) −ψ t . ( ) 2 = Varh2* 6.0. ψ t ( Dt′PD Substituting t t) (33).1772  .

W. This means that the random factors affecting inflation include but do not equal to the ones of risky assets’ prices. (1969). Mathematical Finance. 387-406. 366. 10.2005.org/10. 215-236. Mathematical Finance. http://dx. Markowitz. H.org/10.doi. 19. In our model. In addition. P.doi. This means that the inflation plays as a penalty factor for portfolio revision. J. The Journal of Finance. M. The Journal of Finance. & Zhou. 11371362 and 71173216). we have provided a closed-form optimal strategy and efficient frontier.doi. (1952). Merton. H. http://dx. Continuous-Time Mean-Variance Portfolio Selection with Bankruptcy Prohibition. Y. 1201-1238. Y. S. http://dx. 7. Dual Method for Continuous-Time Markowitz’s Problems with Nonlinear Wealth Equations. Journal of Economic 27 .00218. J. Acknowledgements We acknowledge the contributions of Fundamental Research Funds for the Central Universities (2012QNB19) and Natural Science Foundation of China (11101422.doi. 15. References Bensoussan. X. The driving factors of inflation are not the same ones which affect risky assets’ prices. Dynamic Asset Allocation under Inflation.2010.1111/1467-9965..00362.1016/j. R. Xu. (2005).00459 Ji. 77-91. it tells us that the impact of it cannot be ignored in the real world when portfolio managers choose investment strategy.org/10.org/10. Wu Figure 1. the inflation process is assumed to be a geometric Brownian motion. 213-244. Review of Economics and Statistics.0960-1627.. Optimal Consumption and Portfolio Decisions with Partially Observed Real Prices. H. (1971).1111/j. http://dx. 5.044 Li. R. Furthermore. & Sethi... W. Z.01. (2009). a numerical example is also provided. It can be seen that the frontier with inflation lies below the one without inflation. T.1467-9965. 57. C. we compare our model with that of Zhou & Li (2000).00100 Markowitz.1111/j. our results in this paper are more general. L. & Ng. (1959). Portfolio Selection: Efficient Diversification of Investments.jmaa. Y. By using the more general stochastic LQ approach. R. The biggest difference is that our model is considered the factor of inflation in the decision-making process.Y. 51. (2002).doi. Lifetime Portfolio Selection under Uncertainty: The Continuous-Time Case.doi.. Mathematical Finance.org/10. Keppo.x Brennan.. (2000). The efficient frontier with and without inflation. C. http://dx. S. Journal of Mathematical Analysis and Applications. Next.2307/1926560 Merton. Portfolio Selection. which is correlated with those risky assets. (2010). The search on the liability and bankruptcy prohibition in this problem is left for future work. http://dx. 247-257. Optimum Consumption and Portfolio Rules in a Continuous-Time Model. R. Figure 1 shows the efficient frontier to continuous-time mean-variance model with and without inflation in a market.x Bielecki. Conclusion This paper extends the work of Zhou & Li (2000) to an incomplete market with inflation.1111/1540-6261. Optimal Dynamic Portfolio Selection: Multiperiod Mean-Variance Formulation.2009. D. A. New York: Wiley. Pliska. S.org/10. Comparing to Zhou & Li (2000). 90-100. & Xia.. Jin.

org/10.1016/j. Y. & Li. 943-953. Applied Mathematics and Optimization.2007. J. Y.doi. D. (2000). http://dx. Stochastic Controls: Hamiltonian Systems and HJB Equations (Vol. & Wang. Li. X. 43). 42. http://dx. 42.10. S. Z.014 Yong. 3. 373-413. W. & Zhou. Z. Insurance: Mathematics and Economics.. S.org/10.insmatheco..doi. New York: Springer.doi.org/10.1007/978-1-4612-1466-3 Zhou.1007/s002450010003 28 . (1999). Wu Theory.org/10. http://dx.1016/0022-0531(71)90038-X Xie. X. http://dx. Xu.Y. Continuous-Time Portfolio Selection with Liability: Mean-Variance Model and Stochastic LQ Approach... 19-33. (2008).doi. Y. Continuous-Time Mean-Variance Portfolio Selection: A Stochastic LQ Framework.