Abstract
This paper solves the problem of optimal dynamic investment, consumption and healthcare spending with isoelastic utility, when natural mortality grows exponentially to reflect the Gompertz law and investment opportunities are constant. Healthcare slows the natural growth of mortality, indirectly increasing utility from consumption through longer lifetimes. Optimal consumption and healthcare imply an endogenous mortality law that is asymptotically exponential in the old-age limit, with lower growth rate than natural mortality. Healthcare spending steadily increases with age, both in absolute terms and relative to total spending. The optimal stochastic control problem reduces to a nonlinear ordinary differential equation with a unique solution, which has an explicit expression in the old-age limit. The main results are obtained through a novel version of Perron’s method.
Similar content being viewed by others
Notes
In their words, geographical differences in life expectancy for individuals in the lowest income quartile were significantly correlated with health behaviours such as smoking […], but were not significantly correlated with access to medical care, physical environmental factors, income inequality, or labor market conditions.
References
Bayraktar, E., Zhang, Y.: Minimizing the probability of lifetime ruin under ambiguity aversion. SIAM J. Control Optim. 53, 58–90 (2015)
Beeler, J., Campbell, J.Y.: The long-run risks model and aggregate asset prices: an empirical assessment. Crit. Finance Rev. 1, 141–182 (2012)
Bommier, A.: Portfolio choice under uncertain lifetime. J. Public Econ. Theory 12, 57–73 (2010)
Bommier, A., Rochet, J.-C.: Risk aversion and planning horizons. J. Eur. Econ. Assoc. 4, 708–734 (2006)
Chetty, R., Stepner, M., Abraham, S., Lin, S., Scuderi, B., Turner, N., Bergeron, A., Cutler, D.: The association between income and life expectancy in the United States, 2001–2014. JAMA 315, 1750–1766 (2016)
Cohen, S.N., Elliott, R.J.: Stochastic Calculus and Applications, 2nd edn. Springer, Cham (2015)
Crandall, M.G., Ishii, H., Lions, P.-L.: User’s guide to viscosity solutions of second order partial differential equations. Bull. Am. Math. Soc. 27, 1–67 (1992)
Cutler, D., Deaton, A., Lleras-Muney, A.: The determinants of mortality. J. Econ. Perspect. 20(3), 97–120 (2006)
Ehrlich, I.: Uncertain lifetime, life protection, and the value of life saving. J. Health Econ. 19, 341–367 (2000)
Ehrlich, I., Chuma, H.: A model of the demand for longevity and the value of life extension. J. Polit. Econ. 98, 761–782 (1990)
Gompertz, B.: On the nature of the function expressive of the law of human mortality, and on a new mode of determining the value of life contingencies. Philos. Trans. R. Soc. Lond. 115, 513–583 (1825)
Grossman, M.: On the concept of health capital and the demand for health. J. Polit. Econ. 80, 223–255 (1972)
Hall, R.E., Jones, C.I.: The value of life and the rise in health spending. Q. J. Econ. 122, 39–72 (2007)
Harrison, G.W., Lau, M.I., Rutström, E.E.: Estimating risk attitudes in Denmark: a field experiment. Scand. J. Econ. 109, 341–368 (2007)
Hartman, M., Catlin, A., Lassman, D., Cylus, J., Heffler, S.: US health spending by age, selected years through 2004. Health Aff. 27(1), w1–w12 (2008)
Huang, H., Milevsky, M.A., Salisbury, T.S.: Optimal retirement consumption with a stochastic force of mortality. Insur. Math. Econ. 51, 282–291 (2012)
Hugonnier, J., Pelgrin, F., St-Amour, P.: Health and (other) asset holdings. Rev. Econ. Stud. 80, 663–710 (2013)
Janeček, K., Sîrbu, M.: Optimal investment with high-watermark performance fee. SIAM J. Control Optim. 50, 790–819 (2012)
Richard, S.F.: Optimal consumption, portfolio and life insurance rules for an uncertain lived individual in a continuous time model. J. Financ. Econ. 2, 187–203 (1975)
Rosen, S.: The value of changes in life expectancy. J. Risk Uncertain. 1, 285–304 (1988)
Shepard, D.S., Zeckhauser, R.J.: Survival versus consumption. Manag. Sci. 30, 423–439 (1984)
Smith, J.P.: Healthy bodies and thick wallets: the dual relation between health and economic status. J. Econ. Perspect. 13(2), 145–166 (1999)
Smith, J.P.: The impact of socioeconomic status on health over the life-course. J. Hum. Resour. 42, 739–764 (2007)
Yaari, M.E.: Uncertain lifetime, life insurance, and the theory of the consumer. Rev. Econ. Stud. 32, 137–150 (1965)
Yogo, M.: Portfolio choice in retirement: health risk and the demand for annuities, housing, and risky assets. J. Monet. Econ. 80, 17–34 (2016)
Acknowledgements
We thank for helpful comments from seminar participants at Collegio Carlo Alberto, ETH Zürich, University of Limerick, Alfred Rényi Institute, National Central University in Taiwan, the QMF conference at UTS Sydney, the Congress of the Bachelier Finance Society, the University of Colorado at Boulder, and National Center for Theoretical Sciences in Taiwan. This paper is dedicated to Nicomede Guasoni (1939–2012).
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
P. Guasoni is partially supported by the ERC (278295), NSF (DMS-1412529) and SFI (16/SPP/3347 and 16/IA/4443). Y.-J. Huang is partially supported by NSF (DMS-1715439) and the University of Colorado (11003573).
Appendices
Appendix A: Proofs of main results
In this section, we prove the main results in Sect. 3 by a verification argument, relying on Theorem B.1 and Proposition B.3. Given \(f:\mathbb{R}_{+}\to \mathbb{R}\), consider its Legendre transform \(\widetilde{f}(y):=\sup _{x\ge 0}\{f(x)-xy\}\) for \(y\in \mathbb{R}_{+}\). A heuristic derivation as in Sect. 4.1 shows that the Hamilton–Jacobi equation associated with \(V(x,m)\) in (2.8) is
This is simply (4.9), without the second-order term that stems from the additional risky asset in Sect. 4.
1.1 A.1 Neither aging nor healthcare (\(\beta =0\) and \(g \equiv 0\))
Recall the setup in Sect. 3.1. Since \(V(x,m)\) is nondecreasing in \(x\) by definition, the supremum in the last term of (A.1) vanishes, leading to the equation
If \(V\) is of the form \(V(x,m) = \frac{x^{1-\gamma }}{1-\gamma } v(m)\), the above equation reduces to
where \(c_{0}(m)\) is defined as in (3.4). Setting \(v(m)=u(m)^{- \gamma }\), we obtain from the above equation
whence \(u(m) = c_{0}(m)\). We then prove Proposition 3.1 by verification.
Proof of Proposition 3.1
Set \(w(x,m):=\frac{x^{1- \gamma }}{1-\gamma } c_{0}(m)^{-\gamma }\). Note that (3.3) implies that \(c_{0}(m) >0\) for all \(\gamma >0\), \(\gamma \neq 1\).
Case I:\(0<\gamma <1\). By Theorem B.1, it suffices to verify (B.2) and (B.3). For any \(x\ge 0\), \(c\in \mathcal{C}\) and \(n\in \mathbb{N}\), since \(0<\gamma <1\),
where the convergence follows from (3.3). This already verifies (B.2). On the other hand, since \(\tau _{n}\) is the sum of \(n\) independent, identically distributed exponential random variables with mean \(1/m\),
where the third equality requires \(\delta +m+(\gamma -1)r>0\), which is true under (3.3). Since (3.3) also implies \(\frac{m\zeta ^{(1-\gamma )}}{\delta +m+(\gamma -1)r}{\in} (0,1)\), we deduce \(\mathbb{E}[e^{-\delta \tau ^{m}_{n}} w(\zeta ^{n} X^{0,x,c}_{ \tau ^{m}_{n}},m)]{\to} 0\) as \(n\to \infty \), which verifies (B.3).
Case II:\(\gamma >1\). By Proposition B.3, it suffices to establish (B.16)–(B.18) and show that \(\hat{c}_{t} \equiv c_{0}(m)\) satisfies (B.2) and (B.3). For any \(\varepsilon >0\), since \(g\equiv 0\), the sequence \((\tau ^{\varepsilon }_{n})_{n\in \mathbb{N}_{0}}\) constructed in Appendix B coincides with \((\tau _{n})_{n\in \mathbb{N}_{0}}\). The counting process \(N^{\varepsilon }\) in (B.14) is therefore the same as \(N\) in (2.5). It follows that (B.18) trivially holds. Given \(x\ge 0\), \(c\in \mathcal{C}\) and \(\varepsilon >0\), consider
By construction, \(X^{0,x+\varepsilon , c^{\varepsilon }}_{t} = X^{0,x, c} + \varepsilon e^{rt}\) for all \(t\ge 0\). This together with \(\gamma >1\) implies that for any \(n\in \mathbb{N}\),
Since the right-hand side converges to 0 a.s. as \(t\to \infty \), the above inequality in particular implies that (B.16) holds. On the other hand, a calculation similar to (A.2) shows that
By (3.3), \(\frac{m\zeta ^{(1-\gamma )}}{\delta +m+( \gamma -1)r}\in (0,1)\). We then have \(\mathbb{E}[e^{-\delta \tau _{n}} w(\zeta ^{n} X^{0,x,c}_{\tau _{n}},m)]\to 0\) as \(n\to \infty \), which verifies (B.17). Now, with \(\hat{c}_{t} \equiv c_{0}(m)\), for any \(n\in \mathbb{N}\),
Observing that (3.3) implies
we conclude that \(e^{-(\delta +m) (t-\tau _{n})} w(X^{0,x,\hat{c}} _{t},m)\to 0\) a.s. as \(t\to \infty \). This shows that \(\hat{c}\) satisfies (B.2). Thanks to (A.4), a calculation similar to (A.2) yields
which shows that \(\hat{c}\) satisfies (B.3). □
1.2 A.2 Aging without healthcare (\(g\equiv 0\))
Recall the setup in Sect. 3.2. The Hamilton–Jacobi equation (A.1) associated with the value function \(V(x,m)\) now becomes
If \(V\) is of the form \(V(x,m) = \frac{x^{1-\gamma }}{1-\gamma } v(m)\), the above equation reduces to
where \(c_{0}(m)\) is given by (3.4). Setting \(v(m)=u(m)^{- \gamma }\), we obtain from the above equation
which admits the general solution
where \(C\in \mathbb{R}\) is a constant to be determined. Proposition 3.2 states that taking \(C=0\), which turns \(u(m)\) into \(u_{0}(m)\) in (3.6), leads to the value function \(V(x,m)\). In the following, we separate the proof of Proposition 3.2 into two parts.
Lemma A.1
Under the assumptions of Proposition 3.2, \(u_{0}\)defined in (3.6) is a strictly increasing function on\((0,\infty )\)satisfying (a) and (b) in Proposition 3.2.
Proof
The definition of \(u_{0}\) in (3.6) directly implies that \(u_{0}\) is strictly increasing, \(u_{0}(0)= \frac{\delta +(\gamma -1)r}{ \gamma }\) and \(u_{0}'(0+)=\infty \). Since \(u_{0}\) solves (A.5), for any \(m\in (0,\infty )\),
where the inequality follows from \(u_{0}\) being positive and strictly increasing. On the other hand, using \(y+1 < e^{y}\) for \(y > 0\), (3.6) yields
Finally, the asymptotic expansion of \(u_{0}(m)\) at infinity is
This implies \(u_{0}(m) - ( c_{0}(m) +\beta ) \to 0\) as \(m\to \infty \) and
□
Proof of Proposition 3.2
Set \(w(x,m) := \frac{x^{1- \gamma }}{1-\gamma } u_{0}(m)^{-\gamma }\). By Lemma A.1, it remains to show that \(V(x,m) = w(x,m)\) and \(\hat{c}_{t} := u_{0}(m e^{\beta t})\), \(t\ge 0\), is an optimal control for (3.2). First, we observe that \(\hat{c}\) is an element of \(\mathcal{C}\). Indeed, for any compact subset \(K\) of \(\mathbb{R}_{+}\), thanks to \(u_{0} \le c_{0} +\beta \) in Lemma A.1,
Now we deal with two cases separately.
Case I: Condition (i) holds. By Theorem B.1, it suffices to verify (B.2) and (B.3). For any \((x,m)\in \mathbb{R}^{2}_{+}\), \(c\in \mathcal{C}\) and \(n\in \mathbb{N}\), as \(X^{0,x,c}_{t} \le X^{0,x,c}_{\tau _{n}}\exp (r(t-\tau _{n}))\) on the set \(\{t\ge \tau _{n}\}\) and as \(\gamma \in (0,1)\), we get
where the convergence follows from \(\delta + (\gamma -1) r >0\) and \(u_{0}\) being nondecreasing by definition. This in particular implies (B.2). On the other hand,
where the fourth inequality follows from \(\delta + (\gamma -1) r >0\) and the convergence from \(\gamma ,\zeta \in (0,1)\). Thus (B.3) is satisfied.
Case II: Condition (ii) holds. Due to Proposition B.3, it suffices to establish (B.16)–(B.18) and show that \(\hat{c} _{t} := u_{0}(me^{\beta t})\) satisfies (B.2) and (B.3). For any \(\varepsilon >0\), since \(g\equiv 0\), the sequence \((\tau ^{\varepsilon }_{n})_{n\in \mathbb{N} _{0}}\) constructed in Appendix B coincides with \((\tau _{n})_{n\in \mathbb{N}_{0}}\). The counting process \(N^{\varepsilon }\) in (B.14) is therefore the same as \(N\) in (2.5). Thus (B.18) trivially holds in the current context. Given \((x,m)\in \mathbb{R}^{2}_{+}\), \(c\in \mathcal{C}\) and \(\varepsilon >0\), consider the consumption policy \(c^{\varepsilon }\) as in (A.3) and the property \(X^{0,x+ \varepsilon , c^{\varepsilon }}_{t} = X^{0,x,c}_{t} + \varepsilon e ^{rt}\) for all \(t\ge 0\). We then deduce from \(\gamma >1\) and \(u_{0}\) being a nondecreasing function that for any \(n\in \mathbb{N}\),
which in particular implies (B.16). Since \(\gamma >1\) ensures \(\delta +(\gamma -1)r>0\),
where the convergence follows from \(\gamma ,\zeta >1\). This verifies (B.17). Now for any \(n\in \mathbb{N}\), applying \(\hat{c}_{t} := u_{0}(me^{\beta t})\), \(t\ge 0\), yields
on the set \(\{t \ge \tau _{n}\}\). By direct calculation, \(u_{0}(m)\) defined in (3.6) equals
where \(\overline{\varGamma }(s,z):=\int _{z}^{\infty }t^{s-1} e^{-t} \,dt\) is the upper incomplete gamma function. Recalling the property \(\frac{\overline{\varGamma }(s,z)}{e^{-z}z^{s-1}}\to 1\) as \(z\to \infty \), it follows that
whence \(me^{\beta s} > \gamma u_{0}(m e^{\beta s})\) for \(s\) large enough. This together with \(u_{0}\) being a positive nonincreasing function shows that \(\int _{\tau _{n}}^{t} (m e^{\beta s}-(\gamma -1)u _{0}(me^{\beta s})) \,ds\to \infty \) a.s. as \(t\to \infty \). We then conclude from (A.6) that \(\hat{c}\) satisfies (B.2). It remains to show that \(\hat{c}\) satisfies (B.3). Observe that
where the third line follows from \(e^{-(\delta +(\gamma -1)r) \tau _{n}}\le 1\) as \(\gamma >1\). For each \(n\in \mathbb{N}\), note that \(\sum _{i=1}^{n} Z_{i}\) has a gamma distribution with the law \(\mathbb{P}[\sum _{i=1}^{n} Z_{i}\le z] = \frac{1}{\varGamma (n)} \underline{\varGamma}(n, z)\), where \(\varGamma (s):=\int _{0}^{\infty }t^{s-1} e^{-t} \,dt\) is the gamma function and \(\underline{\varGamma}(s,z):=\int _{0}^{z} t ^{s-1} e^{-t} \,dt\) the lower incomplete gamma function. We then observe from (3.5) that
Differentiating \(\mathbb{P}[\tau _{n}\le t]\) with respect to \(t\) gives
Also, for any
(A.7) yields \(\frac{m}{\gamma u_{0}(m)}> \frac{\alpha }{1- \zeta ^{1-\gamma }}\) for \(m\) large enough. Thus there exists \(t^{*}>0\) such that
Setting \(C(t^{*}):=\exp (\int _{0}^{t^{*}} (\gamma -1)u_{0}(me^{\beta s})\,ds)\), we obtain from (A.8) that
where the second line follows from (A.9) and (A.11), the fourth is due to \(y:=e^{\beta t}-1\), and the last equality holds as \(1-\frac{1}{\alpha }(1-\frac{1}{\gamma })(1- \zeta ^{1-\gamma })>0\) thanks to (A.10). Noting that (A.10) gives \(1-\frac{1}{\alpha }(1-\frac{1}{\gamma })(1- \zeta ^{1-\gamma })> 1-(1-\zeta ^{1-\gamma }) =\zeta ^{1-\gamma }\), we conclude from the above inequality that \(\mathbb{E}[e^{-\delta \tau _{n}} w(\zeta ^{n} X^{0,x,\hat{c}}_{\tau _{n}},me^{\beta \tau _{n}})] \to 0\) as \(n\to \infty \), i.e., \(\hat{c}\) satisfies (B.3). □
1.3 A.3 Aging with healthcare
Recall the setup in Sect. 3.3, where mortality increases naturally according to the Gompertz law (\(\beta >0\)) and at the same time healthcare is available (i.e., \(g:\mathbb{R}_{+}\to \mathbb{R}_{+}\) is not constantly 0) to slow down the mortality growth. For the rest of this section, let Assumption 3.3 hold and denote by \(I:\mathbb{R}_{+}\to \mathbb{R} _{+}\) the inverse function of \(g'\). Note that \(I\) is strictly decreasing.
If the value function (2.8) has the form \(V(x,m) = \frac{x ^{1-\gamma }}{1-\gamma } v(m)\), (A.1) yields
where \(c_{0}(m)\) is given by (3.4). By setting \(v(m)=u(m)^{- \gamma }\) and assuming \(u'(m)\ge 0\), the above equation becomes
Since \(g\) is a nondecreasing function with \(g(0)=0\), the infimum above equals 0. That is, when \(\gamma >1\), the above equation reduces to (A.5), and the associated value function and optimal consumption strategy are as described in Proposition 3.2.
In consequence, we focus on the case \(0 < \gamma < 1\)in the rest of this section. Equation (A.12) is now \(\mathcal{L}u(m) = 0\) as in (3.8). In the following, we employ Perron’s method to construct solutions to (3.8) under the assumption that
Definition A.2
Let \(\varPi \) be the collection of \((p,q)\), where \(p,q:\mathbb{R}_{+} \to \mathbb{R}\) are continuous and satisfy
-
1.
\(c_{0}\le p\le q\le c_{0} + \beta \) on \((0,\infty )\).
-
2.
\(p\) and \(q\) are strictly increasing and concave.
-
3.
\(p\) (resp. \(q\)) is a viscosity subsolution (resp. supersolution) to (3.8) on \((0,\infty )\).
For any \((p,q)\in \varPi \), let \(\mathcal{S}(p,q)\) denote the collection of continuous \(f:\mathbb{R}_{+}\to \mathbb{R}\) such that
-
1.
\(p\le f\le q\) on \((0,\infty )\).
-
2.
\(f\) is strictly increasing and concave.
-
3.
\(f\) is a viscosity supersolution to (3.8) on \((0,\infty )\).
Remark A.3
Under (A.13) and (3.7), \(\varPi \neq \emptyset \). Indeed, \(c_{0}+\beta \) is a supersolution to (A.5) and thus a supersolution to (3.8). Specifically, for any \(m>0\),
On the other hand, \(c_{0}\) is a subsolution to (3.8); indeed, for any \(m>0\),
where \(a:= \frac{1-\zeta ^{1-\gamma }}{\gamma }\) and the inequality follows from (3.7). Thus \(\varPi \) contains at least \((c_{0},c _{0} +\beta )\). Also note that for each \((p,q)\in \varPi \), we have \(\mathcal{S}(p,q)\neq \emptyset \) because by construction \(q\in \mathcal{S}(p,q)\).
We first present a basic result for strictly increasing concave functions \(f\) bounded by \(c_{0}\) and \(c_{0}+\beta \). Note that the concavity of \(f\) implies \(f'(\infty ) := \lim _{m\to \infty } f'(m-)\) is well defined.
Lemma A.4
Assume\(0<\gamma <1\)and (A.13). For any nonnegative, strictly increasing and concave\(f:\mathbb{R}_{+}\to \mathbb{R}\), we have\(\frac{f(m)}{m f'(m-)}\ge 1\)for all\(m\in (0,\infty )\). If\(f\)additionally satisfies\(c_{0}\le f\le c_{0}+\beta \)on\((0,\infty )\), then\(f'(\infty ) =\frac{1-\zeta ^{1-\gamma }}{\gamma }\)and\(\frac{f(m)}{mf'(m-)}\to 1\)as\(m\to \infty \).
Proof
With \(f\) being strictly increasing and \(f(0)\ge 0\), the concavity of \(f\) implies \(\frac{f(m)}{m}\ge f'(m-) >0\) and thus \(\frac{f(m)}{m f'(m-)}\ge 1\) for all \(m\in (0,\infty )\). Suppose \(f\) also lies between \(c_{0}\) and \(c_{0}+\beta \). Since \(c_{0}\) is a linear function with slope \(\frac{1-\zeta ^{1-\gamma }}{\gamma }\), if \(f'( \infty ) \neq \frac{1-\zeta ^{1-\gamma }}{\gamma }\), then \(f(m)\notin [c_{0}(m), c_{0}(m)+\beta ]\) for \(m\) large enough, which is a contradiction. Moreover, we deduce from \(c_{0}(m)\le f(m)\le c_{0}(m)+ \beta \) and \(f'(\infty ) = \frac{1-\zeta ^{1-\gamma }}{\gamma }>0\) that
This implies \(\frac{f(m)}{mf'(\infty )}\to 1\) as \(m\to \infty \). □
The next result shows that \(c_{0}+\alpha \) is a supersolution to (3.8) on \((0,\infty )\) for \(\alpha \) large enough.
Lemma A.5
Assume\(0<\gamma <1\), (A.13) and (3.7). For any\(\alpha \in [0,\beta ]\), \(c_{0}+\alpha \)is a supersolution to (3.8) on\((0,\infty )\)if and only if\(\alpha \in [\beta _{g}, \beta ]\), where\(\beta _{g}\)is defined in (3.10). Specifically,
Proof
For any \(a,b>0\), consider the function
where
By direct calculation,
where the positivity follows from \(I'(y) = \frac{d}{dy}(g')^{-1}(y) = \frac{1}{g''((g')^{-1}(y))}= \frac{1}{g''(I(y))}\) and \(g''<0\).
For any \(\alpha \in [0,\beta ]\), \(c_{0}(m)+\alpha = am+b\) with \(a=\frac{1-\zeta ^{1-\gamma }}{\gamma }\) and \(b=\bar{c}+\alpha \). Then (A.14) reduces to
Observe from (A.15) that \(\beta -\ell (m)\to \beta _{g}\) as \(m\to \infty \). It follows that
This already shows that if \(\alpha \in [0,\beta _{g})\), \(c_{0}+\alpha \) cannot be a supersolution to (3.8) on \((0,\infty )\).
It remains to show that if \(\alpha \in [\beta _{g},\beta ]\), \(c_{0}+\alpha \) satisfies \(\mathcal{L}({c_{0}+\alpha })(m)>0\) for all \(m>0\). This is true for \(\alpha = \beta \), as explained in Remark A.3. For any \(\alpha \in (\beta _{g},\beta )\), using \(a=\frac{1-\zeta ^{1-\gamma }}{\gamma }\) and \(b=\bar{c}+\alpha \) in the current setting and (A.15), (A.17) becomes
implying \(\theta '(m)\to \alpha -\beta <0\) as \(m\downarrow 0\). This together with \(\lim _{m\to \infty }\theta (m)=\infty \) in (A.20) and \(\theta ''(\cdot )>0\) in (A.18) shows that \(\theta \) must attain a global minimum at some \(m^{*}\in (0,\infty )\). Using \(\theta '(m^{*})=0\), we obtain from (A.21) that
The global minimum is then computed as
where the second equality comes from (A.15), the third and fourth follow from (A.22), and the final inequality is due to (3.7). We thus conclude that for any \(\alpha \in (\beta _{g}, \beta )\), \(\mathcal{L}(c_{0}+\alpha )(m) = \theta (m) \ge \theta (m ^{*}) >0\) for all \(m\in (0,\infty )\). Finally, for the case \(\alpha =\beta _{g}\), since \(c_{0}+\beta _{g}\) is the pointwise infimum of the supersolutions \(c_{0}+\alpha \), \(\alpha \in (\beta _{g},\beta ]\), it must also be a supersolution thanks to Crandall et al. [7, Lemma 4.2]. Observe from (A.21) that \(\lim _{m\uparrow \infty }\theta '(m) = a( \beta _{g}-\beta _{g}) =0\). This together with \(\theta ''>0\) in (A.18) and \(c_{0}+\beta _{g}\) being a supersolution shows that \(\theta (m)\) must be strictly decreasing on \((0,\infty )\) with \(\lim _{m\uparrow \infty }\theta (m)\ge 0\). As a result, \(\mathcal{L}(c _{0}+\beta _{g})(m)=\theta (m)>0\) for all \(m>0\). □
Lemma A.6
Assume\(0<\gamma <1\), (A.13) and (3.7). For any\((p,q)\in \varPi \), we then have\(p < c_{0}+\beta _{g}\)on\(\mathbb{R} _{+}\), with\(\beta _{g}\)defined in (3.10).
Proof
By contradiction, suppose that ‘\(p < c_{0}+\beta _{g}\) on \(\mathbb{R} _{+}\)’ does not hold. If there exists \(m_{0}>0\) such that \(p(m)= c _{0}(m)+\beta _{g}\) for all \(m\ge m_{0}\), then the subsolution property of \(p\) is violated, thanks to Lemma A.5. Thus it remains to deal with the second case: there exists \(m_{0}>0\) such that \(p(m)> c_{0}(m)+\beta _{g}\) for \(m>m_{0}\).
Since \(p'(\infty )=\frac{1-\zeta ^{1-\gamma }}{\gamma }\) by Lemma A.4 (i), there must exist \(\alpha _{0} \in (\beta _{g},\beta ]\) such that \(p\le c_{0}+\alpha _{0}\) and \((c_{0}(m)+\alpha )-p_{0}(m)\downarrow 0\) as \(m\to \infty \). Consider the collection of functions \(\{c_{0}+\alpha :\alpha \in (\frac{\alpha _{0}+\beta _{g}}{2},\alpha _{0})\}\). For each \(\alpha \in (\frac{\alpha _{0}+\beta _{g}}{2},\alpha _{0})\), we introduce \(\theta ^{\alpha }(m):= \mathcal{L}(c_{0}+\alpha )(m)\) and recall the formula for \(\theta ^{\alpha }\) in (A.19). Then we get
where \(a:=\frac{1-\zeta ^{1-\gamma }}{\gamma }\). Moreover, in view of (A.15) with \(b\) replaced by \(\bar{c}+\alpha \), the above convergence is uniform in \(\alpha \in (\frac{\alpha _{0}+\beta _{g}}{2}, \alpha _{0})\). That is, for any \(\delta >0\), there is \(M(\delta )>0\) such that for all \(m\ge M(\delta )\) and \(\alpha \in (\frac{\alpha _{0}+\beta _{g}}{2},\alpha _{0})\), \(|\frac{\theta ^{\alpha }(m)}{m}- a(\alpha - \beta _{g})|<\delta \). Taking \(\delta := \frac{a(\alpha _{0}-\beta _{g})}{4}\), we get for any \(m>M(\delta )\) and \(\alpha \in (\frac{\alpha _{0}+\beta _{g}}{2},\alpha _{0})\) that
Fix \(\varepsilon \in (0,\frac{a(\alpha _{0}-\beta _{g})}{4\beta })\). We can take \(\alpha \in (\frac{\alpha _{0}+\beta _{g}}{2},\alpha _{0})\) large enough such that \(c_{0}+\alpha \) intersects \(p\) at \(m^{*}> M(\delta )\) and \(a< p'(m^{*})< a+\varepsilon \). It follows that
where the third equality follows from \(\theta ^{\alpha }(m^{*})= \mathcal{L}(c_{0}+\alpha )(m^{*})\) and the last inequality is due to (A.23) and the choice of \(\varepsilon \). This implies that \(p\) cannot be a subsolution to (3.8) on \((0,\infty )\), which is a contradiction. □
Following Perron’s method, we introduce for each \((p,q)\in \varPi \) the function
Proposition A.7
(Supersolution property)
Assume\(0<\gamma <1\)and (A.13). For any\((p,q)\in \varPi \), we have\(u^{*}_{p,q}\in \mathcal{S}(p,q)\).
Proof
As a pointwise infimum of concave nondecreasing functions between \(c_{0}\) and \(c_{0}+\beta \), \(u^{*}_{p,q}\) is by definition concave, nondecreasing and between \(c_{0}\) and \(c_{0}+\beta \). The concavity of \(u^{*}_{p,q}\) yields the desired continuity. Then by Crandall et al. [7, Lemma 4.2], \(u^{*}_{p,q}\), being continuous and a pointwise infimum of viscosity supersolutions, is again a viscosity supersolution. It remains to show that \(u^{*}_{p,q}\) is strictly increasing. Suppose to the contrary that \(u^{*}_{p,q}\equiv \kappa >0\) in a neighbourhood of some \(m^{*}\in (0,\infty )\). The concavity of \(u^{*}_{p,q}\) then implies that \(u^{*}_{p,q}\equiv \kappa \) on \([m^{*},\infty )\). It follows that \(\mathcal{L}u ^{*}_{p,q}(m) = \kappa (\kappa -c_{0}(m))<0\) as \(m\) large enough. This contradicts the supersolution property of \(u^{*}_{p,q}\). □
Proposition A.8
(Subsolution property)
Let\(0<\gamma <1\)and assume (A.13). Fix\((p,q)\in \varPi \). Suppose\(u^{*}_{p,q}\)is strictly concave at\(m_{0}\in (0,\infty )\)in the following sense:
Then\(u^{*}_{p,q}\)is a viscosity subsolution to (3.8) at\(m_{0}\).
Proof
If \(u^{*}_{p,q}\) is strictly concave at \(m_{0}\in (0,\infty )\) as in (A.24), there are three cases:
(i) \((u^{*}_{p,q})'(m_{0} -)\neq (u^{*}_{p,q})'(m_{0}+)\);
(ii) \((u^{*}_{p,q})'(m_{0} -) = (u^{*}_{p,q})'(m_{0}+)\), and \(u^{*}\) is strictly concave on \([m_{0}-\kappa ,m_{0}+\kappa ]\) for some \(\kappa >0\);
(iii) \((u^{*}_{p,q})'(m_{0} -) = (u^{*}_{p,q})'(m_{0}+)\), and there exists \(\kappa _{1},\kappa _{2}>0\) such that \(u^{*}_{p,q}\) is linear on \([m_{0}-\kappa _{1},m_{0}]\) and strictly concave on \([m_{0},m_{0}+ \kappa _{2}]\), or strictly concave on \([m_{0}-\kappa _{1},m_{0}]\) and linear on \([m_{0},m_{0}+\kappa _{2}]\).
Assume by contradiction that there exists a test function \(\psi \in C^{1}((0,\infty ))\) such that \(0 = (u^{*}_{p,q}-\psi )(m_{0}) > (u ^{*}_{p,q}-\psi )(m)\) for all \(m\in (0,\infty )\setminus \{m_{0}\}\) and \(\mathcal{L}\psi (m_{0})>0\). For the cases (i) and (ii), we can assume without loss of generality that \(\psi \) is strictly increasing and concave on \((0,\infty )\). Take \(\delta >0\) small enough such that \(\mathcal{L}\psi (m)>0\) for all \(m\in (m_{0}-\delta ,m_{0}+\delta )\). Then for small enough \(\varepsilon >0\), one can take \(0<\delta _{1} \le \delta \) such that for each \(0<\eta \le \varepsilon \), \(\mathcal{L}(\psi -\eta )(m)>0\) for all \(m\in (m_{0}-\delta _{1},m_{0}+ \delta _{1})\). Consider the function
When \(\eta \) is small enough, \(u^{\eta }\) is by construction a concave, strictly increasing viscosity supersolution to (3.8) on \((0,\infty )\), and \(u^{*}_{p,q}-\eta \le u^{\eta }\le u^{*}_{p,q}\). That is, \(u^{\eta }\in \mathcal{S}(p,q)\) when \(\eta \) is small enough. However, by definition, \(u^{\eta }< u^{*}_{p,q}\) in some small neighbourhood of \(m_{0}\), which contradicts the definition of \(u^{*}_{p,q}\).
Now we deal with the case (iii). Set \(a:= (u^{*}_{p,q})'(m_{0} -) = (u ^{*}_{p,q})'(m_{0}+)\). In view of (3.8), to get the desired subsolution property, it suffices to prove
We assume without loss of generality that \(u^{*}_{p,q}\) is linear on \([m_{0}-\kappa _{1},m_{0}]\) and strictly concave on \([m_{0},m_{0}+ \kappa _{2}]\). Take \((\ell _{n})_{n\in \mathbb{N}}\) in \((m_{0},m_{0}+ \kappa _{2}]\) such that \(\ell _{n}\downarrow m_{0}\) and \(u^{*}_{p,q}\) is differentiable at \(\ell _{n}\). Then the subsolution property established above for case (ii) implies that \(\mathcal{L} u^{*}_{p,q}(\ell _{n}) \le 0\) for all \(n\in \mathbb{N}\). Observe that the map
as \(g\) is strictly concave, nondecreasing and \(g'(\infty )=0\). By the continuity of \(u^{*}_{p,q}\) and (A.26), \(\mathcal{L} u ^{*}_{p,q}(\ell _{n}) \le 0\) implies (A.25) as \(n\to \infty \). □
We next establish the strict concavity of \(u^{*}_{p,q}\). Recall that \(I\) denotes the inverse function of \(g'\).
Proposition A.9
(Strict concavity)
Assume\(0<\gamma <1\), (A.13) and (3.7). For any\((p,q)\in \varPi \), \(u^{*}_{p,q}\)is strictly concave on\((0,\infty )\).
Proof
Assume by contradiction that \(u^{*}_{p,q}\) is linear, i.e., \(u^{*}_{p,q}(m) =am+b\), on some interval of \(\mathbb{R}_{+}\). Since \(u^{*}_{p,q}\in \mathcal{S}(p,q)\), we deduce from Lemma A.4 that \(a\ge \frac{1-\zeta ^{1-\gamma }}{ \gamma }\) and \(b\in [\bar{c}, \bar{c}+\beta ]\). Recall \(\theta (m):= \mathcal{L}(am+b)\) from (A.14).
Case I:\(a=\frac{1-\zeta ^{1-\gamma }}{\gamma }\) and \(b\in [\bar{c},\bar{c}+\beta _{g})\): Then we have \(u^{*}_{p,q}(m) = c _{0}(m)+\alpha \) for \(m\) large enough, where \(\alpha := b-\bar{c} \in [0,\beta _{g})\). Thanks to Lemma A.5, we have \(\lim _{m\uparrow \infty }\mathcal{L}(u^{*}_{p,q})(m) = \lim _{m\uparrow \infty }\mathcal{L}(c_{0}+\alpha )(m)=-\infty \). This contradicts the supermartingale property of \(u^{*}_{p,q}\).
Case II:\(a=\frac{1-\zeta ^{1-\gamma }}{\gamma }\) and \(b\in [\bar{c}+\beta _{g}, \bar{c}+\beta ]\): There are three sub-cases.
Case II-1:\(u^{*}_{p,q}(m) = am+b\) for all \(m\ge 0\), with \(b\in (\bar{c}+\beta _{g}, \bar{c}+\beta ]\): Let us write \(u^{*}_{p,q}(m) = c_{0}(m)+\alpha \) with \(\alpha := b-\bar{c}\in (\beta _{g},\beta ]\). For any \(\bar{\alpha }\in (\beta _{g},\alpha )\), Lemmas A.5 and A.6 imply that \(c_{0}+\bar{\alpha }\) belongs to \(\mathcal{S}({p,q})\) and is strictly less than \(u^{*}_{p,q}\), which contradicts the definition of \(u^{*}_{p,q}\).
Case II-2:\(u^{*}_{p,q}(m) = am+(\bar{c} +\beta _{g})\) for all \(m\ge 0\): By (A.14) and (A.21), \(\lim _{m\downarrow 0}\theta (m)= b(b-\bar{c})>0\) and \(\lim _{m\downarrow 0}\theta '(m)= b- \bar{c}-\beta <0\). Thus we can take \(m^{*}>0\) small enough such that \(\theta (m^{*})>0\) and \(\theta '(m^{*})<0\). In view of the continuous dependence of \(\theta (m^{*})\) and \(\theta '(m^{*})\) on \(a,b\) in (A.14) and (A.17), there exists \(\delta > 0\) small enough such that when \(a,b\) are replaced by \(\bar{a} \in (a,a+\delta )\) and \(\bar{b} \in (b-\delta ,b)\), \(\theta (m^{*})>0\) and \(\theta '(m^{*})<0\) still hold. Take suitable \(\bar{a} \in (a,a+\delta )\) and \(\bar{b} \in (b-\delta ,b)\) such that \(\bar{a} m^{*}+\bar{b} = u^{*}_{p,q}(m ^{*})\) and \(\bar{a} m +\bar{b} > p(m)\) for \(m\in (0,m^{*}]\) (such \(\bar{a}\) and \(\bar{b}\) exist by Lemma A.6). For clarity, let \(\bar{\theta }\) and \(\bar{\theta }'\) denote \(\theta \) and \(\theta '\) with \(a,b\) replaced by \(\bar{a},\bar{b}\). Now we deduce from \(\lim _{m\downarrow 0}\bar{\theta }(m)=\bar{b}(\bar{b}-\bar{c})>0\) (obtained from (A.14) as above), \(\bar{\theta }(m^{*})>0\), \(\bar{\theta }'(m^{*})<0\) and \(\bar{\theta }''(m)>0\) for all \(m> 0\) (by (A.18)) that \(\bar{\theta }(m)>0\) for all \(m\in (0, m^{*})\). Consider the function \(\psi (m):= \bar{a} m+ \bar{b}\). By definition, \(\mathcal{L}\psi (m)=\bar{\theta }(m)>0\) for \(m\in (0,m^{*})\). Thus \(\psi \wedge u^{*}_{p,q}\) belongs to \(\mathcal{S}(p,q)\) and is strictly less than \(u^{*}_{p,q}\) for \(m\in (0, m^{*})\). This contradicts the definition of \(u^{*}_{p,q}\).
Case II-3: There exists \(m_{0}>0\) such that \(u^{*}_{p,q}(m) = am+b\) for all \(m\ge m_{0}\) and \(u^{*}_{p,q}\) is strictly concave at \(m_{0}\) in the sense of (A.24): By Proposition A.8, \(u^{*}_{p,q}\) is a viscosity subsolution to (3.8) at \(m_{0}\). Consider the test function \(\psi (m) := am+ b\), \(m\in (0,\infty )\), of \(u^{*}_{p,q}\) at \(m_{0}\). The subsolution property of \(u^{*}_{p,q}\) yields \(\mathcal{L}\psi (m_{1})\le 0\). Note that \(\psi (m) = c_{0}(m)+\alpha \) with \(\alpha :=b-\bar{c}\in [\beta _{g},\beta ]\). Thus by Lemma A.5, \(\mathcal{L}{\psi }(m)>0\) for all \(m>0\), which is a contradiction.
Case III:\(a>\frac{1-\zeta ^{1-\gamma }}{\gamma }\) and \(b=\bar{c}\): There exists \(m_{0}>0\) with the property that \(u ^{*}_{p,q}(m) = am+b\) for \(m\in [0,m_{0}]\) and \(u^{*}_{p,q}\) is strictly concave at \(m_{0}\) in the sense of (A.24). By Proposition A.8, \(u^{*}_{p,q}\) is a viscosity subsolution to (3.8) at \(m_{0}\). For all \(m\in (0,\infty )\), define
with \(\ell \) as in (A.15). Note that \(\theta (m) = m\eta (m)\). By direct calculation and (A.16),
Since we currently have \(b = \bar{c}\), \(\eta '(m) >0\) for all \(m\in (0,\infty )\). Take \(\psi (m) := am+b\), \(m\in (0,\infty )\), as a test function of \(u^{*}_{p,q}\) at \(m_{0}\). The subsolution property of \(u^{*}_{p,q}\) implies \(0\ge \mathcal{L}\psi (m_{0})=\theta (m_{0})=m _{0}\eta (m_{0})\). We therefore have \(\eta (m)<0\) for all \(m\in (0,m _{0})\). The supersolution property of \(u^{*}_{p,q}\), however, entails \(0\le \mathcal{L}u^{*}_{p,q}(m)=\theta (m)=m\eta (m)\) for all \(m\in (0,m_{0})\), which is a contradiction.
Case IV:\(a> \frac{1-\zeta ^{1-\gamma }}{\gamma }\) and \(b\in (\bar{c}, \bar{c}+\beta )\): There are two sub-cases.
Case IV-1: There exists \(m_{0}>0\) such that \(u^{*}_{p,q}(m) = am+b\) for all \(m\in [0,m_{0}]\) and \(u^{*}_{p,q}\) is strictly concave at \(m_{0}\) in the sense of (A.24): We first show that \(p(0)\) has to be strictly less than \(u^{*}_{p,q}(0)\). If \(p(0)= u^{*}_{p,q}(0)\), then \(\lim _{m\downarrow 0}p'(m) \le a\); otherwise, \(p(m)>u^{*}_{p,q}(m)\) for \(m> 0\) small enough, which contradicts \(u^{*}_{p,q}\in \mathcal{S}(p,q)\). By the concavity of \(p\), we can take a real sequence \((\ell _{n})\) such that \(\ell _{n} \downarrow 0\) and \(p\) is differentiable at \(\ell _{n}\). The subsolution property of \(p\) then implies \(\mathcal{L}p(\ell _{n})\le 0\) for all \(n\in \mathbb{N}\). As \(n\to \infty \), we get \(p(0)(p(0)-\bar{c}) \le 0\), thanks to the finiteness of \(\lim _{m\downarrow 0}p'(m)\). This shows that \(p(0)<\bar{c}\), contradicting \(p\ge c_{0}\).
By Proposition A.8, \(u^{*}_{p,q}\) is a viscosity subsolution to (3.8) at \(m_{0}\). Consider the test function \(\psi (m) := am+b\), \(m\in (0,\infty )\), of \(u^{*}_{p,q}\) at \(m_{0}\). The subsolution property of \(u^{*}_{p,q}\) implies \(0\ge \mathcal{L}\psi (m_{0})=\theta (m_{0})\). By (A.14), \(\lim _{m\downarrow 0}\theta (m)= b(b-\bar{c})>0\). If \(\lim _{m\downarrow 0}\theta '(m)\ge 0\), then \(\theta ''>0\) on \((0,\infty )\) (by (A.18)) implies that \(\theta (m)>\theta (0)>0\) for all \(m>0\), which contradicts \(\theta (m_{0})\le 0\). If \(\lim _{m\downarrow 0} \theta '(m)< 0\), we can follow the argument in Case II-2. Take \(0< m^{*}< m_{0}\) small enough such that \(\theta (m^{*})>0\) and \(\theta '(m^{*})<0\). By the continuous dependence of \(\theta (m^{*})\) and \(\theta '(m^{*})\) on \(a,b\), there exists \(\delta > 0\) such that when \(a,b\) are replaced by \(\bar{a} \in (a,a+\delta )\) and \(\bar{b} \in (b- \delta ,b)\), \(\theta (m^{*})>0\) and \(\theta '(m^{*})<0\) still hold. Choose suitable \(\bar{a} \in (a,a+\delta )\) and \(\bar{b} \in (b- \delta ,b)\) such that \(\bar{a} m^{*}+\bar{b} = u^{*}_{p,q}(m^{*})\) and \(\bar{a} m +\bar{b} > p(m)\) for \(m\in (0,m^{*}]\) (such \(\bar{a}\) and \(\bar{b}\) exist because \(p(0)< u^{*}_{p,q}(0)\)). For clarity, let \(\bar{\theta }\) and \(\bar{\theta }'\) denote \(\theta \) and \(\theta '\) with \(a,b\) replaced by \(\bar{a},\bar{b}\). Now we deduce from \(\lim _{m\downarrow 0}\bar{\theta }(m)>0\), \(\bar{\theta }(m^{*})>0\), \(\bar{\theta }'(m^{*})<0\) and \(\bar{\theta }''>0\) on \((0,\infty )\) that \(\bar{\theta }(m)>0\) for all \(m\in (0, m^{*})\). Consider the function \(\phi (m):= \bar{a} m+\bar{b}\). By definition, \(\mathcal{L}\phi (m)=\bar{ \theta }(m)>0\) for \(m\in (0,m^{*})\). Thus \(\phi \wedge u^{*}_{p,q}\) belongs to \(\mathcal{S}(p,q)\) and is strictly less than \(u^{*}_{p,q}\) for \(m\in (0, m^{*})\). This contradicts the definition of \(u^{*}_{p,q}\).
Case IV-2: There are \(m_{1}, m_{2}\in (0,\infty )\) with the property that \(u^{*}_{p,q}(m) = am+b\) for \(m\in [m_{1},m_{2}]\) and \(u^{*}_{p,q}\) is strictly concave at \(m_{1}\) and \(m_{2}\) in the sense of (A.24): By Proposition A.8, \(u^{*}_{p,q}\) is a viscosity subsolution to (3.8) at both \(m_{1}\) and \(m_{2}\). Take \(\psi (m) := am+b\), \(m\in (0,\infty )\), as a test function of \(u^{*}_{p,q}\) at \(m_{1}\) and \(m_{2}\). Then the subsolution property of \(u^{*}_{p,q}\) implies \(0\ge \mathcal{L}\psi (m_{1})=\theta (m_{1})\) and \(0\ge \mathcal{L}\psi (m_{2})=\theta (m_{2})\). As \(\theta ''>0\) on \((0,\infty )\) by (A.18), we must have \(\theta (m_{3})<0\) for some \(m_{3}\in (m_{1},m_{2})\). The supersolution property of \(u^{*}_{p,q}\), however, entails \(0\le \mathcal{L}\psi (m)=\theta (m)\) for all \(m\in (m_{1},m_{2})\), which is a contradiction. □
Proposition A.10
(Regularity)
Let\(0<\gamma <1\)and assume (A.13) and (3.7). For any\((p,q)\in \varPi \), \(u^{*}_{p,q}\)is a strictly concave classical solution to (3.8) on\((0,\infty )\).
Proof
For any \((p,q)\in \varPi \), Propositions A.7–A.9 imply that \(u^{*}_{p,q}\) is a strictly concave viscosity solution to (3.8) on \((0,\infty )\). It remains to show that \(u^{*}_{p,q}\) is differentiable everywhere on \((0,\infty )\). Assume by contradiction that there exists \(m_{0}\in (0,\infty )\) such that \(a:=(u^{*})'(m_{0}+) < (u^{*})'(m_{0}-)=:b\). Take \((k_{n})_{n\in \mathbb{N}}\) and \((\ell _{n})_{n\in \mathbb{N}}\) in \((0,\infty )\) such that \(k_{n} \uparrow m_{0}\), \(\ell _{n}\downarrow m_{0}\) and \(u^{*}_{p,q}\) is differentiable at \(k_{n}\) and \(\ell _{n}\) for all \(n\in \mathbb{N}\). By the viscosity solution property of \(u^{*}_{p,q}\), \(\mathcal{L}u^{*}(k _{n}) = \mathcal{L}u^{*}(\ell _{n})=0\) for all \(n\in \mathbb{N}\). As \(n\to \infty \), it follows that
which implies that \(a=b\), a contradiction. □
Proposition A.11
(Verification)
Let\(0 < \gamma < 1\)and suppose (A.13) and (3.7) hold. If\(u:\mathbb{R}_{+}\to \mathbb{R}_{+}\)is a nonnegative, strictly increasing and concave classical solution to (3.8) on\((0,\infty )\), then
Furthermore, \((\hat{c},\hat{h})\)defined by
is an optimal control for (2.8).
Proof
Set \(w(x,m) := \frac{x^{1-\gamma }}{1-\gamma } u(m)^{-\gamma }\). In view of Theorem B.1, it suffices to show that (B.2) and (B.3) hold and \((\hat{c},\hat{h})\) belongs to \(\mathcal{A}\). Since \(u\) is nonnegative, strictly increasing and concave, Lemma A.4 implies that \(\hat{h} _{t} \le I(\frac{1-\gamma }{\gamma })\) for all \(t\ge 0\). Moreover, there exist \(a,b>0\) such that \(u(m)< am+b\) for all \(m\ge 0\). It follows that for any compact subset \(K\subset \mathbb{R}_{+}\),
This already shows that \((\hat{c},\hat{h})\in \mathcal{A}\).
As \(u\) is a classical solution to (3.8), (3.7) implies \(u^{2}(m) - u(m) c_{0}(m) \ge 0\) and thus \(u(m)\ge c_{0}(m)\) for all \(m\in (0,\infty )\). Now, given \((x,m)\in \mathbb{R}^{2}_{+}\), \((c,h)\in \mathcal{A}\) and \(n\in \mathbb{N}\), using \(0<\gamma <1\) and \(X^{0,x,c,h}_{t} \le X^{0,x,c,h}_{\tau _{n}}\exp \left (r(t-\tau _{n})\right )\) yields
where the second line is due to \(M^{0,m,h}_{t} \ge 0\) by definition, the third follows from \(u\) being strictly increasing with \(u(m)> u(0) \ge c_{0}(0)=\bar{c}>0\), and the convergence is a consequence of \(\delta + (\gamma -1) r = \gamma \bar{c}>0\). This in particular implies (B.2). On the other hand, for each \(n\in \mathbb{N}\),
where the last inequality is due to \(\delta +(\gamma -1)r =\gamma \bar{c}>0\). This shows that (B.3) also holds. □
Proposition A.11 together with Propositions A.10 and A.7 leads to
Corollary A.12
Assume\(0 < \gamma < 1\), (A.13) and (3.7). Then\(u^{*}_{p,q}\)is independent of the choice of\((p,q)\in \varPi \), and it is the unique nonnegative, strictly increasing and concave classical solution to (3.8) on\((0,\infty )\).
Remark A.13
Proposition A.11 and Corollary A.12 yield Theorem 3.4.
In the following, we simply denote by \(u^{*}\) the function \(u^{*}_{p,q}\) for any \((p,q) \in \varPi \).
Corollary A.14
(Strict concavity of \(u_{0}\))
Assume\(0 < \gamma < 1\)and (A.13). Then\(u_{0}\)defined in (3.6) is strictly concave on\((0,\infty )\).
Proof
With \(g\equiv 0\), Eq. (3.8) reduces to (A.5). We can repeat the same arguments as in this section (with much simpler proofs) to show that the strictly concave \(u^{*}\) constructed via Perron’s method coincides with \(u_{0}\). □
We are now ready to prove Theorem 3.5.
Proof of Theorem 3.5
Observe that \(\beta _{g} = \beta - g(I(\frac{1- \gamma }{\gamma })) + \frac{1-\gamma }{\gamma } I(\frac{1-\gamma }{ \gamma })\). Then (3.7) implies \(\beta _{g}>0\). Since \(u_{0}\) is a solution to (A.5) and \(u_{0}'(m)\ge 0\), it is a supersolution to (3.8). This together with Lemma A.1, Corollary A.14 and Remark A.3 shows that \((c_{0},u_{0})\in \varPi \). As a result, \(u^{*} = u^{*}_{c_{0},u_{0}}\le u_{0}\). Similarly, \((c_{0},c _{0}+\beta _{g})\in \varPi \) by Lemma A.5, which implies \(u^{*}=u^{*}_{c_{0},c_{0}+\beta _{g}}\le c_{0}+\beta \). This already yields the relation \(u^{*}\le \min \{u_{0}, c_{0}+\beta _{g}\}\). On the other hand, thanks again to Lemma A.1 and Corollary A.14 with \(\beta \) replaced by \(\beta _{g}\), \(u_{0}^{g}\) is nonnegative, strictly increasing, concave and bounded from below and above by \(c_{0}\) and \(c_{0}+ \beta _{g}\), respectively. Then Lemma A.4 implies \(\frac{u^{g}_{0}(m)}{m (u ^{g}_{0})'(m)}\ge 1\) for all \(m>0\). It follows that
Since \(u_{0}^{g}\) is by construction a solution to (A.5) with \(\beta \) replaced by \(\beta _{g}\), the above inequality gives
for all \(m>0\). This shows \((u^{g}_{0},c_{0}+\beta _{g})\in \varPi \), whence \(u^{*}=u^{*}_{u^{g}_{0},c_{0}+\beta _{g}}\ge u^{g}_{0}\). □
Appendix B: Verification
In this section, we provide a general verification theorem for the value function \(V(x,m)\) in (2.8). Given \((c,h)\in \mathcal{A}\), we introduce for each \(n\in \mathbb{N}\) the truncated policies \((c^{(n)},h^{(n)})\in \mathcal{A}\) as
Theorem B.1
Let\(w\in C^{1,1}(\mathbb{R}_{+}\times \mathbb{R}_{+})\)satisfy (A.1). Suppose for any\((x,m)\in \mathbb{R}^{2}_{+}\)and\((c,h)\in \mathcal{A}\)that
Then:
(i) \(w(x,m) \ge V(x,m)\)on\(\mathbb{R}_{+}\times \mathbb{R} _{+}\).
(ii) Suppose there exist two measurable functions\(\bar{c}\), \(\bar{h}:\mathbb{R}^{2}_{+}\to \mathbb{R}_{+}\)such that\(\bar{c}(x,m)\)and\(\bar{h}(x,m)\)are maximisers of
respectively, for all\((x,m)\in \mathbb{R}^{2}_{+}\). Let\(\bar{X}\), \(\bar{M}\), \(\bar{N}\)denote the solutions to
where\(T_{0}:=0\)and\(T_{n+1}:= \inf \{t\ge T_{n} : \int _{T_{n}}^{t} M _{s} \,ds\ge Z_{n+1}\}\)for\(n\ge 1\). Define the processes\((\hat{c}, \hat{h})\)by
If\((\hat{c},\hat{h})\in \mathcal{A}\), then\((\hat{c},\hat{h})\)is an optimal control for (2.8) and\(w(x,m)=V(x,m)\)on\(\mathbb{R}_{+}\times \mathbb{R}_{+}\).
Proof
(i) Given \((c,h)\in \mathcal{A}\), recall that
for some \((c_{n})\), \((h_{n})\in \mathfrak{L}\). We claim that for all \(n\in \mathbb{N}\), we have
First, we prove this for \(n=1\). Since \(c_{0}\) and \(h_{0}\) are deterministic functions and \(w\) is a classical solution to (A.1),
for all \(t\ge 0\). Letting \(t\to \infty \) and in view of (B.2),
Thanks to Fubini’s theorem and (2.6), observe that
whence (B.5) holds true for \(n=1\) in view of (B.7)–(B.9). Now suppose (B.5) holds true for \(n=k>1\). That is,
By writing \(x_{k} = \zeta ^{k}X^{0,x,c,h}_{\tau _{k}}\) and \(m_{k} = M ^{0,m,h}_{\tau _{k}}\), we get (B.6) with \((0,x,m,c_{0},h_{0})\) replaced by \((\tau _{k},x_{k},m_{k},c_{k},h_{k})\). This together with (B.2) gives
Using Fubini’s theorem and (2.6) as in (B.8), (B.9), the above inequality implies that
This together with (B.10) shows that
The claim (B.5) then holds by induction. As \(n\to \infty \) in (B.5), monotone convergence and (B.3) yield \(w(x,m)\ge \mathbb{E}[\int _{0}^{\infty } e^{-\delta t} U(c_{t} \zeta ^{N_{t}}{X}^{0,x,c,h}_{t}) \,dt]\) for all \((c,h)\in \mathcal{A}\). Taking the supremum over \((c,h)\in \mathcal{A}\) gives \(w(x,m)\ge V(x,m)\).
(ii) With \((c,h)=(\hat{c},\hat{h})\), the inequality (B.6) turns into an equality, whence (B.5) holds with equality. As \(n\to \infty \), monotone convergence and (B.3) imply
This together with part (i) shows that \(w(x,m)=V(x,m)\) and \((\hat{c}, \hat{h})\) is an optimal control. □
Theorem B.1 is now extended to include the risky asset \(S\) in (4.1). Recall the setup in Sect. 4, especially \(\mathcal{A}'\) in (4.4) and the value function \(V\) in (4.5). For any \((c,h,\pi )\in \mathcal{A}'\), we can also consider the truncated version \((c^{(n)}, h^{(n)}, \pi ^{(n)}) \in \mathcal{A}'\) defined as in (B.1).
Theorem B.2
Let\(w\in C^{1,2}(\mathbb{R}_{+}\times \mathbb{R}_{+})\)satisfy (4.7). Suppose for any\((x,m)\in \mathbb{R}^{2}_{+}\)and\((c,h,\pi )\in \mathcal{A}'\)that (B.2) and (B.3) hold, where\(X^{0,x,c^{(n)},h^{(n)}}\)is replaced by\(X^{0,x,c^{(n)},h^{(n)},\pi ^{(n)}}\)and\(\mathbb{E}[\ \cdot \mid Z _{1},\dots ,Z_{n}]\)by\(\mathbb{E}[\ \cdot \mid \mathcal{F}_{\tau _{n}}]\). Then:
(i) \(w(x,m) \ge V(x,m)\)on\(\mathbb{R}_{+}\times \mathbb{R} _{+}\).
(ii) Suppose there exist measurable functions\(\bar{c}\), \(\bar{h}\), \(\bar{\pi }:\mathbb{R}^{2}_{+}\to \mathbb{R}_{+}\)with\(\bar{c}\)and\(\bar{h}\)as described in Theorem B.1 (ii) and\(\bar{\pi }(x,m)\)being the maximiser of
Let\(\bar{X}\), \(\bar{M}\), \(\bar{N}\)denote the solutions to
where\(T_{0}:=0\)and\(T_{n+1}:= \inf \{t\ge T_{n}: \int _{T_{n}}^{t} M _{s} \,ds\ge Z_{n+1}\}\)for\(n\ge 1\). Consider\((\hat{c},\hat{h}, \hat{\pi })\)with\(\hat{c}\)and\(\hat{h}\)defined as in (B.4) and\(\hat{\pi }_{t}:=\bar{\pi }(\zeta ^{\bar{N}_{t}} \bar{X}_{t}, \bar{M}_{t})\)for\(t\ge 0\). If\((\hat{c},\hat{h}, \hat{\pi })\)is in\(\mathcal{A}'\), defined in (4.4), then\((\hat{c},\hat{h}, \hat{\pi })\)is an optimal control for (4.5) and\(w(x,m)=V(x,m)\)on\(\mathbb{R}_{+} \times \mathbb{R}_{+}\).
Proof
(i) We follow the arguments in Theorem B.1, replacing \(X^{0,x,c,h}\) by \(X^{0,x,c,h,\pi }\) in (4.2). For any \((\hat{c}, \hat{h},\hat{\pi })\in \mathcal{A}'\), we now prove (B.5) for all \(n\in \mathbb{N}\). For \(n=1\), as \(w\) is a solution to (4.7), Itô’s formula yields (B.6), with the left-hand side and the second line under the expectation \(\mathbb{E} _{2}\). By (B.2), letting \(t\to \infty \) gives (B.7), with the right-hand side under the expectation \(\mathbb{E}_{2}\). Fubini’s theorem and (4.3) imply (B.8) and (B.9), with their second lines again under the expectation \(\mathbb{E}_{2}\). Thus (B.5) holds for \(n=1\). Now suppose (B.5) holds for \(n=k>1\), i.e., (B.10) is true. As \(w\) is a solution to (4.7) and in view of (B.2), by Itô’s formula, (B.11) holds with \(\mathbb{E}[\ \cdot \mid Z_{1},\dots ,Z_{n}]\) replaced by \(\mathbb{E}[\ \cdot \mid \mathcal{F}_{\tau _{n}}]\). By Fubini’s theorem and (4.3) as above, (B.12) follows. This together with (B.10) implies that (B.5) holds for \(n=k+1\). Thus (B.5) follows by induction. Letting \(n\to \infty \) in (B.5) and recalling (B.3), the same argument as at the end of the proof of Theorem B.1 (i) yields that \(w(x,m)\ge V(x,m)\).
(ii) This follows from the same argument as in Theorem B.1 (ii). □
In the sequel, we relax the conditions in Theorem B.1. For any \((x,m)\in \mathbb{R}^{2}_{+}\) and \((c,h)\in \mathcal{A}\), suppose that the household is given additional wealth \(\varepsilon >0\) at time 0 and decides not to spend it at all over time. Imagine that at time 0, the household deposits \(x\ge 0\) in a standard account and \(\varepsilon >0\) in a separate additional account. Then the household behaves as if there was no additional wealth: at each time \(t\ge 0\), the amount it spends in consumption (resp. healthcare) is \(c_{t}\) (resp. \(h_{t}\)) multiplied by the standard account balance. The rates of spending in consumption and healthcare therefore become
This new process \(h^{\varepsilon }\) of spending rate in healthcare, different from \(h\), changes the moments of deaths. More precisely, starting from time 0, the household takes
as instantaneous spending rates in healthcare. As in (2.4), the time of the first death is defined as
Starting from time \(\tau ^{\varepsilon }_{1}\), the household takes
as instantaneous spending rates in healthcare. Set \(m^{\varepsilon } _{1}:= M^{0,m,h^{\varepsilon }_{0}}_{\tau ^{\varepsilon }_{1}}\); then the time of the second death is defined as in (2.4) by
In general, for each \(n\in \mathbb{N}\), the household, starting from time \(\tau ^{\varepsilon }_{n}\), takes
as instantaneous spending rates in healthcare. Set \(m^{\varepsilon } _{n}:= M^{\tau ^{\varepsilon }_{n-1},m^{\varepsilon }_{n-1},h^{\varepsilon }_{n-1}}_{\tau ^{\varepsilon }_{n}}\); then the \((n+1)\)th death moment is defined as in (2.4) by
As in (2.5), we can introduce the counting process
Similarly, define for each \(n\in \mathbb{N}\)
Observe that \((c_{n}^{\varepsilon })\), \((h_{n}^{\varepsilon })\in \mathfrak{L}\), and it can be checked that
This in particular shows that \(( c^{\varepsilon },h^{\varepsilon }) \in \mathcal{A}\). In view of (B.13), we have the identity
Proposition B.3
Let\(w\in C^{1,1}(\mathbb{R}_{+}\times \mathbb{R}_{+})\)satisfy (A.1). Suppose for any\((x,m)\in \mathbb{R}^{2}_{+}\), \((c,h)\in \mathcal{A}\)and\(\varepsilon >0\)that
Then:
(i) \(w(x,m)\ge V(x,m)\)on\(\mathbb{R}_{+}\times \mathbb{R} _{+}\).
(ii) Suppose the measurable functions\(\bar{h}\)and\(\bar{c}\)specified in Theorem B.1 (ii) exist so that we can define\((\hat{c},\hat{h})\)as in (B.4). If\((\hat{c},\hat{h})\)is in\(\mathcal{A}\)and satisfies (B.2) and (B.3), then\((\hat{c},\hat{h})\)is an optimal control for (2.8) and\(w(x,m)= V(x,m)\)on\(\mathbb{R}_{+}\times \mathbb{R}_{+}\).
Proof
We carry out the same arguments as in Theorem B.1. With the aid of (B.16), we obtain (B.5), with \((x,c,h,\tau _{n},N_{t})\) replaced by \((x+\varepsilon ,c^{\varepsilon },h^{\varepsilon },\tau ^{\varepsilon }_{n},N^{\varepsilon }_{t})\). Letting \(n\to \infty \) and using (B.17), we get
where the equality follows from (B.15) and the definition of \(c^{\varepsilon }\) in (B.13). Letting \(\varepsilon \to 0\), we obtain from (B.18) that \(w(x,m)\ge \mathbb{E}[ \int _{0}^{\infty } e^{-\delta t} U(c_{t} \zeta ^{N_{t}} {X}^{0,x,c,h} _{t})\,dt]\). Taking the supremum over \((c,h)\in \mathcal{A}\) leads to \(w(x,m)\ge V(x,m)\). The proof of (ii) is the same as Theorem B.1 (ii). □
Theorem B.2 can also be relaxed in a similar fashion.
Proposition B.4
Let\(w\in C^{1,2}(\mathbb{R}_{+}\times \mathbb{R}_{+})\)satisfy (4.7). Suppose for any\((x,m)\in \mathbb{R}^{2}_{+}\), \((c,h,\pi )\in \mathcal{A}'\)and\(\varepsilon >0\)that (B.16)–(B.18) hold with\(X^{0,x,c_{n}^{\varepsilon },h_{n}^{\varepsilon }}\)replaced by\(X^{0,x,c_{n}^{\varepsilon },h _{n}^{\varepsilon },\pi _{n}^{\varepsilon }}\)and\(\mathbb{E}[\ \cdot \mid Z_{1},\dots ,Z_{n}]\)by\(\mathbb{E}[\ \cdot \mid \mathcal{F}_{ \tau _{n}}]\). Then:
(i) \(w(x,m)\ge V(x,m)\)on\(\mathbb{R}_{+}\times \mathbb{R} _{+}\).
(ii) Suppose the measurable functions\((\bar{c}, \bar{h}, \bar{ \pi })\)specified in Theorem B.2 (ii) exist so that we can define\((\hat{c},\hat{h}, \hat{\pi })\)therein. If\((\hat{c},\hat{h},\hat{\pi })\)is in\(\mathcal{A}'\)and satisfies (B.2) and (B.3) as specified in Theorem B.2, then\((\hat{c},\hat{h}, \hat{\pi })\)is an optimal control for the problem (4.5) and\(w(x,m)= V(x,m)\)on\(\mathbb{R}_{+} \times \mathbb{R}_{+}\).
Rights and permissions
About this article
Cite this article
Guasoni, P., Huang, YJ. Consumption, investment and healthcare with aging. Finance Stoch 23, 313–358 (2019). https://doi.org/10.1007/s00780-019-00383-6
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00780-019-00383-6