数学物理学报  2016, Vol. 36 Issue (1): 187-192   PDF (347 KB)    
扩展功能
加入收藏夹
复制引文信息
加入引用管理器
Email Alert
RSS
本文作者相关文章
刘晓
陈振龙
带注资的对偶模型中征税问题
刘晓1,2, 陈振龙1    
1 浙江工商大学统计与数学学院 杭州 310018;
2 安徽师范大学数学计算机科学学院 安徽 芜湖 241003
摘要: 该文在带注资的对偶模型中研究征税问题.假设税按照loss-carry-forward制度支付.当盈余低于0时,将采取注资的方式使得盈余达到0而不致破产.假设收益服从指数分布,得到了期望折现征税总额减去期望折现注资成本总额的显式表达式.
关键词: 征税     对偶模型     注资    
Taxation Problems in the Dual Model with Capital Injections
Liu Xiao1,2 , Chen Zhenlong1    
1 School of Statistics and Mathematics, Zhejiang Gongshang University, Hangzhou 310018;
2 School of Mathematics and Computer Science, Anhui Normal University, Anhui Wuhu 241003
Abstract: This paper studies the taxation problems in the dual model with capital injections. Assume that the taxes are paid according to a loss-carry-forward system. Let the surplus attain 0 avoiding ruin by capital injections when the surplus becomes negative. The explicit expression for the expected discounted tax payments minus the expected discounted costs of capital injections is obtained under the assumption that the gains are exponentially distributed.
Key words: Taxation     Dual model     Capital injection    

1 引言

假设本文引入的所有随机过程和随机变量都定义在完备概率空间$(\Omega,{\cal F},\{{\cal F}_{t}\}_{t\geq 0},P)$上,盈余过程$\{U(t);t\geq 0\}$由对偶模型描述,即 \begin{equation} U(t)=x-ct+\sum_{i=1}^{N(t)}Y_{i}, \tag{1.1}\end{equation} 其中$U(0)=x\geq 0$是初始盈余,$c>0$为均匀费用率, $\{N(t);t\geq 0\}$是强度为$\lambda>0$的泊松过程, $\{Y_{i};i=1,2,\cdots\}$是独立同分布的指数收益随机变量序列, 共同分布函数为$F(y)=1-e^{-\beta y},y>0$. 假设$\{N(t);t\geq 0\}$与$\{Y_{i};i=1,2,\cdots\}$相互独立.对$i=1,2,\cdots$,记$W_{i}$为第$i$次收益到达时刻, $T_{i}=W_{i}-W_{i-1}$,其中$W_{0}=0$,则$\{T_{i};i=1,2,\cdots\}$ 是独立同分布的随机变量序列,共同分布是均值为$\frac{1}{\lambda}$的指数分布. 模型(1.1)适用于制药和石油勘探类的公司,因为这类公司平时会连续地投入一些成本, 如人员工资和研究费用,但一旦研究成功,会有一些突然的收益. 对偶模型近年来受到大量精算研究者的关注,如Avanzi等[1], Liu和Chen[2],刘东海和刘再明[3],Peng等[4], 袁海丽和胡亦钧[5]等.

本文考虑固定税率征税策略. 在每一次收益到达时刻,如果该次收益到达一个新的历史盈余最高点, 那么征税数额为该次最高点与上次最高点差额部分的$\gamma(0\leq\gamma<1)$倍.这种征税制度称为loss-carry-forward制度. Loss-carry-forward征税制度首先由Albrecher和Hipp[6]提出,近年来,有大量文献在不同的风险模型中, 基于loss-carry-forward制度,研究破产概率、征税现值和最优征税策略等问题[7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18], 其中Albrecher等[7]在模型(1.1)中研究了破产概率、破产时间的Laplace变换、折现征税随机变量和提高征税起点时最优征税起点问题. 本文在Albrecher等[7]的基础上考虑注资,也就是说,当盈余为负时, 公司立即通过融资使盈余达到0以防止破产,但公司需要支付$k(k>1)$倍的融资成本.记$L(t)(L(0)=0)$为右连续且左极限存在的非降过程,表示到$t$时刻累积注资总额. 记$U^{L}(t)=U(t)+L(t)$. 记$\{S^{U^{L}}(t);t\geq 0\}$表示 $\{U^{L}(t);t\geq 0\}$的流动最大值过程, 即$S^{U^{L}}(t)=\max\limits _{0\leq s\leq t}U^{L}(s)$. 记征税后的盈余过程为$\{U_{\gamma}^{L}(t);t\geq 0\}$,则 \begin{equation} U_{\gamma}^{L}(t)=U^{L}(t)-\gamma(S^{U^{L}}(t)-x).\tag{1.2}\end{equation}

我们的目的是要研究期望折现征税总额减去期望折现注资成本总额,即 \begin{equation} u(x)=E_{x}\bigg{[}\frac{\gamma}{1-\gamma}\sum_{s\geq 0}e^{-\delta s} I_{\{U_{\gamma}^{L}(s)=M_{\gamma}^{L}(s)\}}(M_{\gamma}^{L}(s)-M_{\gamma}^{L}(s-))-k\int_{0}^{\infty}e^{-\delta t}{\rm d}L(t)\bigg{]}, \tag{1.3}\end{equation} 其中$E_{x}[\cdot]$表示相应于初始盈余为$x$的数学期望,$\delta$表示折现率, $M_{\gamma}^{L}(s)=\max\limits _{0\leq r\leq s}U_{\gamma}^{L}(r)$. 在实际应用中, 当公司发生破产时,国家将会通过一定的资金支持维持公司的运营, $u(x)$可以看作公司对国家的净贡献. 关于注资的想法是由Dickson和Waters[19]提出, 并广泛地被引入到各种模型中进行研究[20, 21, 22, 23, 24, 25, 26, 27].

本文结构安排如下: 第二节给出$u(x)$的显式表达式; 第三节考虑提高征税起点的情形, 得到了相应的期望折现征税总额减去期望折现注资成本总额的表达式与未提高征税起点情况下的关系.

2 $u(x)$的显式表达式

对$0\leq x\leq y$,记$g(x,y)$表示过程$\{U^{L}(t);t\geq 0\}$从$x$出发首次超过$y$时注资的期望折现值,$\tau_{x,y}$表示过程$\{U^{L}(t);t\geq 0\}$ 从$x$出发首次超过$y$的时刻,$h(x,y)=E_{x}[e^{-\delta\tau_{x,y}}]$.

引理 2.1  对$0\leq x\leq y$,有 \begin{equation} g(x,y)=\frac{(\beta+r)e^{r(y-x)}-(\beta+s)e^{s(y-x)}}{r(\beta+r)e^{ry}-s(\beta+s)e^{sy}}, \tag{2.1}\end{equation} \begin{equation} h(x,y)=\frac{\lambda}{c}\frac{re^{rx}-se^{sx}}{r(\beta+r)e^{ry}- s(\beta+s)e^{sy}}e^{(r+s)(y-x)}, \tag{2.2}\end{equation} 其中$r,s(r>0>s)$是方程 $$ c\xi^{2}-(\lambda+\delta-\beta c)\xi-\beta\delta=0 $$ 的两个根.

  对$0\leq x\leq y$,考虑经典风险模型 $X(t)=x+ct-\sum\limits_{i=1}^{N(t)}Y_{i}$, 其中$x$,$c$,$\{N(t);t\geq 0\}$和$\{Y_{i};i=1,2,\cdots\}$的解释同模型(1.1). 记$V(x,y)$为以参数为$y$的barrier策略分红(即盈余一旦超过$y$, 超过的部分立即作为分红支出)时,破产前折现分红随机变量的数学期望, $\{X^{y}(t);t\geq 0\}$表示分红后的盈余过程, $T_{x,y}={\rm inf}\{t\geq0:X^{y}(t)<0\}$为破产时刻, $D(x,y)=E_{x}[e^{-\delta T_{x,y}}]$为相应破产时刻的Laplace变换. 根据文献[28, 29]知 $$ V(x,y)=\frac{(\beta+r)e^{rx}-(\beta+s)e^{sx}}{r(\beta+r)e^{ry}-s(\beta+s)e^{sy}}, $$ $$ D(x,y)=\frac{\lambda}{c}\frac{re^{ry+sx}-se^{rx+sy}}{r(\beta+r)e^{ry}-s(\beta+s)e^{sy}}.$$ 通过$\{U^{L}(t);t\geq 0\}$和$\{X^{y}(t);t\geq 0\}$样本轨道的对比, 可以得到$g(x,y)=V(y-x,y)$和$h(x,y)=D(y-x,y)$,从而(2.1)和(2.2)式成立. 证毕.

引理 2.2  对任意$x\geq0$,有$u(x)\leq\frac{\lambda\gamma}{\beta\delta}$.

  显然有 $$u(x)\leq \gamma E_{x}\bigg{[}\sum_{k=1}^{\infty}e^{-\delta\sum\limits_{i=1}^{k}T_{i}}Y_{k}\bigg{]} =\gamma E_{x}(Y_{1}) \sum_{k=1}^{\infty}\Big{[}E_{x}\big{(}e^{-\delta T_{1}}\big{)} \Big{]}^{k}=\frac{\lambda\gamma}{\beta\delta}.$$ 证毕.

记$g(x)=g(x,x)$和$h(x)=h(x,x)$,则 $$ g(x)=\frac{r-s}{r(\beta+r)e^{rx}-s(\beta+s)e^{sx}}, $$ $$ h(x)=\frac{\lambda}{c}\frac{re^{rx}-se^{sx}}{r(\beta+r)e^{rx}-s(\beta+s)e^{sx}}.$$

定理 2.3 对$x\geq0$,有 \begin{equation} u(x)=-h(x)\int_{x}^{\infty}\bigg{[}k\frac{g(y)h'(y)-h(y)g'(y)}{h(y)^{2}} +\frac{\beta kg(y)-\gamma h(y)}{(1-\gamma)h(y)}\bigg{]} e^{-\frac{\beta}{1-\gamma}\int_{x}^{y}[1-h(z)]{\rm d}z}{\rm d}y.\tag{2.3}\end{equation}

 考虑到,当盈余过程从$x$出发,首次超过$x$时, 超出的量(征税后)服从均值为$\frac{1-\gamma}{\beta}$的指数分布, 且注资成本的期望折现值为$kg(x)$,得到 \begin{eqnarray} u(x)&=&-kg(x)+h(x)\bigg{[}\frac{\gamma}{\beta}+\int_{0}^{\infty} u(x+(1-\gamma)y){\rm d}F(y)\bigg{]}\nonumber\\ &=&-kg(x)+h(x)\bigg{[}\frac{\gamma}{\beta}+\frac{\beta}{1-\gamma} e^{\frac{\beta}{1-\gamma}x}\int_{x}^{\infty}u(z) e^{-\frac{\beta}{1-\gamma}z}{\rm d}z\bigg{]}. \tag{2.4}\end{eqnarray} 对(2.4)式两边关于$x$求导,得到 \begin{eqnarray} u'(x)&=&-kg'(x)+h'(x)\bigg{[}\frac{\gamma}{\beta}+\frac{\beta}{1-\gamma} e^{\frac{\beta}{1-\gamma}x}\int_{x}^{\infty}u(z) e^{-\frac{\beta}{1-\gamma}z}{\rm d}z\bigg{]}\nonumber\\ && +h(x)\bigg{[}\frac{\beta^{2}}{(1-\gamma)^{2}}e^{\frac{\beta}{1-\gamma}x} \int_{x}^{\infty}u(z)e^{-\frac{\beta}{1-\gamma}z}{\rm d}z -\frac{\beta}{1-\gamma}u(x)\bigg{]}\nonumber\\ &=&-kg'(x)+\frac{h'(x)}{h(x)}[u(x)+kg(x)]+\frac{\beta}{1-\gamma} \bigg{[}u(x)+kg(x) -\frac{\gamma}{\beta}h(x)-h(x)u(x)\bigg{]}\nonumber\\ &=&\bigg{[}\frac{h'(x)}{h(x)}+\frac{\beta}{1-\gamma}[1-h(x)]\bigg{]} u(x)+k\frac{g(x)h'(x)-h(x)g'(x)}{h(x)}+\frac{\beta kg(x)-\gamma h(x)}{1-\gamma}.~~ \tag{2.5}\end{eqnarray} 微分方程(2.5)的解可表示为 \begin{eqnarray} u(x)&=&h(x)e^{\frac{\beta}{1-\gamma}\int_{0}^{x}[1-h(z)]{\rm d}z} \nonumber\\ &&\cdot \Big{[}\int_{0}^{x}\Big{[}k\frac{g(y)h'(y)-h(y)g'(y)}{h(y)^{2}} +\frac{\beta kg(y)-\gamma h(y)}{(1-\gamma)h(y)}\Big{]} e^{-\frac{\beta}{1-\gamma}\int_{0}^{y}[1-h(z)]{\rm d}z}{\rm d}y+C\Big{]}, \tag{2.6}\end{eqnarray} 其中$C$为某常数. 注意到 \begin{eqnarray*} \lim\limits_{x\rightarrow\infty}h(x)&=&\frac{\lambda}{c(\beta+r)}\\ &=&\frac{2\lambda}{\lambda+\delta+\beta c+\sqrt{(\lambda+\delta-\beta c)^{2} +4\beta c\delta}}\\ &<&\frac{2\lambda}{\lambda+\delta+\beta c+\mid\lambda+\delta-\beta c\mid}\\ &=&\frac{2\lambda}{2\max \{\lambda+\delta,\beta c\}}\\ &<&\frac{\lambda}{\lambda+\delta}, \end{eqnarray*} 有$\lim\limits_{x\rightarrow\infty}e^{\frac{\beta}{1-\gamma}\int_{0}^{x}[1-h(z)]{\rm d}z} =\infty$. 再根据(2.6)式和引理2.2,得到 \begin{eqnarray*} C=-\int_{0}^{\infty}\Big{[}k\frac{g(y)h'(y)-h(y)g'(y)}{h(y)^{2}} +\frac{\beta kg(y)-\gamma h(y)}{(1-\gamma)h(y)}\Big{]} e^{-\frac{\beta}{1-\gamma}\int_{0}^{y}[1-h(z)]{\rm d}z}{\rm d}y.\end{eqnarray*} 将$C$的值代入到(2.6)式即得到(2.3)式. 证毕.

注 2.4  引理2.2中等式能否达到? 从直观上看,可以猜想, 当盈余过程只有上跳而没有任何下降过程时,所征税额应该是最大的, 此时等式可以达到,也就是$c=0$的情形. 下面就来验证一下$c\rightarrow 0$时(2.3)式的极限情形. 注意到,当$c\rightarrow0$时,有$r\rightarrow\infty$且$cr\rightarrow\lambda+\delta$, 从而对$x>0$,有$h(x)\rightarrow\frac{\lambda}{\lambda+\delta}$,$h'(x)\rightarrow0$, $g(x)\rightarrow0$且$g'(x)\rightarrow0$,由(2.3)式知 \begin{eqnarray*} u(x)\rightarrow\frac{\lambda\gamma}{(\lambda+\delta)(1-\gamma)}\int_{x}^{\infty}e^{-\frac{\beta\delta}{(\lambda+\delta)(1-\gamma)}(y-x)}{\rm d}y=\frac{\lambda\gamma}{\beta\delta}, \end{eqnarray*} 从而验证了猜想.

3 提高征税起点

本节将研究提高征税起点时,期望折现征税总额减去期望折现注资成本总额的表达式. 提高征税起点是指,给定一个水平$b>x$, 只有当盈余超过$b$时才开始按照loss-carry-forward制度征税, 记$u(x,b)$为相应的期望折现征税总额减去期望折现注资成本总额. 定理3.1将给出$u(x,b)$与$u(b)$之间的关系.

定理 3.1 对$0\leq x <b$, 有 \begin{equation} u(x,b)=\frac{ru(b)+ke^{-rb}}{re^{-sb}-se^{-rb}}e^{-sx}-\frac{su(b)+ke^{-sb}}{re^{-sb}-se^{-rb}}e^{-rx}.\tag{3.1} \end{equation}

 对首次逸出时间${\tau _{x,b}}$作条件期望, 由盈余过程的强马氏性知 \begin{equation} u(x,b)=-kg(x,b)+h(x,b)\Big{[}\frac{\gamma}{\beta}+\int_{0}^{\infty}u(b+(1-\gamma)y){d}F(y)\Big{]}. \tag{3.2} \end{equation} 由(2.4)式知 \begin{equation} \frac{\gamma}{\beta}+\int_{0}^{\infty}u(b+(1-\gamma)y){d}F(y)=\frac{u(b)+kg(b)}{h(b)}. \tag{3.3} \end{equation} 由引理2.1知 \begin{equation} \frac{h(x,b)}{h(b)}=\frac{re^{rx}-se^{sx}}{re^{rb}-se^{sb}}e^{(r+s)(b-x)}. \tag{3.4} \end{equation} 将(3.3)式代入到(3.2)式, 并根据(3.4)式及g(x,b)和g(b)的表达式, 得到 \begin{eqnarray*} u(x,b)&=&-kg(x,b)+h(x,b)\frac{u(b)+kg(b)}{h(b)}\\ &=&-k\frac{(\beta+r)e^{r(b-x)}-(\beta+s)e^{s(b-x)}}{r(\beta+r)e^{rb}-s(\beta+s)e^{sb}}\\ &&+k\frac{re^{rx}-se^{sx}}{re^{rb}-se^{sb}}\frac{r-s}{r(\beta+r)e^{rb}-s(\beta+s)e^{sb}}e^{(r+s)(b-x)}\\ &&+\frac{re^{rx}-se^{sx}}{re^{rb}-se^{sb}}u(b)e^{(r+s)(b-x)}\\ &=&\frac{re^{rx}-se^{sx}}{re^{rb}-se^{sb}}u(b)e^{(r+s)(b-x)}\\ && -\frac{k[r(\beta+r)e^{r(2b-x)}+s(\beta+s)e^{s(2b-x)}-r(\beta+r)e^{(r+s)b-sx}-s(\beta+s)e^{(r+s)b-rx}]}{[r(\beta+r)e^{rb}-s(\beta+s)e^{sb}][re^{rb}-se^{sb}]}\\ &=&\frac{re^{rx}-se^{sx}}{re^{rb}-se^{sb}}u(b)e^{(r+s)(b-x)} -k\frac{e^{r(b-x)}-e^{s(b-x)}}{re^{rb}-se^{sb}}\\ &=&\frac{ru(b)+ke^{-rb}}{re^{-sb}-se^{-rb}}e^{-sx}-\frac{su(b)+ke^{-sb}}{re^{-sb}-se^{-rb}}e^{-rx}, \end{eqnarray*} 即(3.1)式成立. 证毕.

参考文献
[1] Avanzi B, Gerber H U, Shiu E S W. Optimal dividends in the dual model. Insurance:Mathematics and Economics, 2007, 41:111-123
[2] Liu X, Chen Z L. Dividend problems in the dual model with diffusion and exponentially distributed observation time. Statistics and Probability Letters, 2014, 87:175-183
[3] 刘东海, 刘再明. 线性红利界下的对偶风险模型. 应用数学学报, 2013, 36:1118-1126 Liu D H, Liu Z M. On a dual model with a linear dividend barrier. Acta Mathematicae Applicatae Sinica (Chinese Series), 2013, 36:1118-1126
[4] Peng D, Liu D H, Liu Z M. Dividend problems in the dual risk model with exponentially distributed observation time. Statistics and Probability Letters, 2013, 83:841-849
[5] 袁海丽, 胡亦钧. 带利率和常数红利边界的对偶风险模型的研究. 数学学报, 2012, 55:132-140 Yuan H L, Hu Y J. On the dual risk model with interest and constant dividend barrier. Acta Mathematica Sinica (Chinese Series), 2012, 55:132-140
[6] Albrecher H, Hipp C. Lundberg's risk process with tax. Blätter DGVFM, 2007, 28:13-28
[7] Albrecher H, Badescuc A, Landriaultd D. On the dual risk model with tax payments. Insurance:Mathematics and Economics, 2008, 42:1086-1094
[8] Ming R X, Wang W Y, Xiao L Q. On the time value of absolute ruin with tax. Insurance:Mathematics and Economics, 2010, 46:67-84
[9] Wang W Y, Hu Y J. Optimal loss-carry-forward taxation for the Lévy risk model. Insurance:Mathematics and Economics, 2012, 50:121-130
[10] Wei J Q, Yang H L, Wang R M. On the Markov-modulated insurance risk model with tax. Blätter der DGVFM, 2010, 31:65-78
[11] Albrecher H, Borst S, Boxma O, Resing J. The tax identity in risk theory-a simple proof and an extension. Insurance:Mathematics and Economics, 2009, 44:304-306
[12] Albrecher H, Renaud J F, Zhou X W. A Lévy insurance risk process with tax. Journal of Applied Probability, 2008, 45:363-375
[13] Wang W Y, Ming R X, Hu Y J. On the expected discounted penalty function for risk process with tax. Statistics and Probability Letters, 2011, 81:489-501
[14] Wei L. Ruin probability in the presence of interest earnings and tax payments. Insurance:Mathematics and Economics, 2009, 45:133-138
[15] Kyprianou A E, Zhou X W. General tax structures and the Lévy insurance risk model. Journal of Applied Probability, 2009, 46:1146-1156
[16] Renaud J F. The distribution of tax payments in a Lévy insurance risk model with a surplus-dependent taxation structure. Insurance:Mathematics and Economics, 2009, 45:242-246
[17] Cheung E, Landriault D. On a risk model with surplus dependent premium and tax rates. Methodology and Computing in Applied Probability, 2010, 14:233-251
[18] Dong H, Liu Z M. On a risk model with Markovian arrivals and tax. Applied Mathematics-A Journal of Chinese Universities, 2012, 27:150-158
[19] Dickson D C M, Waters H R. Some optimal dividends problems. Astin Bulletin, 2004, 34:49-74
[20] Bai L H, Paulsen J. On non-trivial barrier solutions of the dividend problem for a diffusion under constant and proportional transaction costs. Stochastic Processes and Their Applications, 2012, 122:4005-4027
[21] 陈树敏, 何春雄. 带比例及固定费用的对偶模型分红策略. 应用概率统计, 2013, 29:136-150 Chen S M, He C X. Optimal dividend phyments in a dual risk model with both fixed and proportional costs. Chinese Journal of Applied Probability, 2013, 29:136-150
[22] Dai H S, Liu Z M, Luan N N. Optimal dividend strategies in a dual model with capital injections. Mathematical Methods of Operations Research, 2010, 72:129-143
[23] Kulenko N, Schimidli H. An optimal dividend strategy in a Cramér-Lundberg model with capital injections. Insurance:Mathmatics and Economics, 2008, 43:270-278
[24] Peng X F, Chen M, Guo J Y. Optimal dividend and equity issuance problem with proportional and fixed transaction costs. Insurance:Mathematics and Economics, 2012, 51:576-585
[25] Yao D J, Wang R M, Xu L. Optimal dividend, capital injection and reinsurance strategies with variance premium principle. Scientia Sinica Mathematica, 2014, 44:1123-1140
[26] Yao D J, Wang R M, Xu L. Optimal dividend and capital injection strategy with fixed costs and restricted dividend rate for a dual model. Journal of Industrial and Management Optimization, 2014, 10:1235-1259
[27] Zhang S Q, Liu G X. Optimal dividend payment and capital injection of the compound Poisson risk model with both proportional and fixed costs. Scientia Sinica Mathematica, 2012, 42:827-843
[28] Schmidli H. Stochastic Control in Insurance. London:Springer-Verlag, 2008
[29] Irbäck J. Asymptotic theory for a risk process with a high dividend barrier. Scandinavian Actuarial Journal, 2003, 2:97-118