P191 命题2

This commit is contained in:
2025-10-14 14:16:12 +08:00
parent 39a4bf0afa
commit 2a0eea1d0a

View File

@ -21,6 +21,7 @@
\max_{\mathbf{h}} \ \mathbf{h}^T \boldsymbol{\alpha} - \frac{\lambda}{2} \mathbf{h}^T V \mathbf{h} \max_{\mathbf{h}} \ \mathbf{h}^T \boldsymbol{\alpha} - \frac{\lambda}{2} \mathbf{h}^T V \mathbf{h}
\] \]
这里不考虑预算约束(可通过无风险资产借贷调节),\(\lambda > 0\) 是风险厌恶系数。 这里不考虑预算约束(可通过无风险资产借贷调节),\(\lambda > 0\) 是风险厌恶系数。
(关于此优化问题详细信息参考第6章)
--- ---
@ -146,8 +147,9 @@ E[R_n] = R_F + \kappa \cdot \mathrm{Cov}(r_n, r_Q) \tag{8A-24}
随机折现因子 \(v(s)\) 满足: 随机折现因子 \(v(s)\) 满足:
\[ \[
E[v R_n] = R_F \quad (8\text{-}17) E[v R_n] = (1+i_F) = R_F \quad (8\text{-}17)
\] \]
且 \(E[v] = 1\)。 且 \(E[v] = 1\)。
由 (8A-25) 由 (8A-25)
@ -173,7 +175,7 @@ E\left[ R_n \cdot \left( 1 - \kappa (r_Q - f_Q) \right) \right] = R_F \quad (8A\
--- ---
## **3. 与 SDF 公式比较** ## **推导 \(v(s)\) 的形式**
由 (8-17) 由 (8-17)
@ -187,10 +189,6 @@ E\left[ R_n \cdot \left( 1 - \kappa (r_Q - f_Q) \right) \right] = E[v R_n]
\] \]
这对任意资产 \(n\) 都成立。 这对任意资产 \(n\) 都成立。
---
## **4. 推导 \(v(s)\) 的形式**
要使: 要使:
\[ \[
@ -210,67 +208,4 @@ v(s) = 1 - \kappa (r_Q(s) - f_Q) + \varepsilon(s)
v(s) = 1 - \kappa (r_Q(s) - f_Q) \quad (8A\text{-}21) v(s) = 1 - \kappa (r_Q(s) - f_Q) \quad (8A\text{-}21)
\] \]
---
## **5. 确定 \(\kappa\)**
由 \(E[v] = 1\)
\[
E[1 - \kappa (r_Q - f_Q)] = 1
\]
即:
\[
1 - \kappa (E[r_Q] - f_Q) = 1
\]
但 \(E[r_Q] = f_Q\),所以自动满足,无法确定 \(\kappa\)。
实际上 \(\kappa\) 由 \(v\) 与 \(R_Q\) 的关系进一步确定。
由 \(E[v R_Q] = R_F\)
\[
E\left[ (1 - \kappa (r_Q - f_Q)) R_Q \right] = R_F
\]
代入 \(R_Q = r_Q + R_F\)
\[
E\left[ (1 - \kappa (r_Q - f_Q)) (r_Q + R_F) \right] = R_F
\]
展开:
\[
E[r_Q + R_F - \kappa r_Q (r_Q + R_F) + \kappa f_Q (r_Q + R_F)] = R_F
\]
利用 \(E[r_Q] = f_Q\),整理得:
\[
f_Q + R_F - \kappa E[r_Q^2] - \kappa R_F f_Q + \kappa f_Q^2 + \kappa R_F f_Q = R_F
\]
\[
f_Q + R_F - \kappa E[r_Q^2] + \kappa f_Q^2 = R_F
\]
\[
f_Q - \kappa (E[r_Q^2] - f_Q^2) = 0
\]
\[
f_Q - \kappa \sigma_Q^2 = 0
\]
所以:
\[
\kappa = \frac{f_Q}{\sigma_Q^2} \quad (8A\text{-}22)
\]
---
**最终**
\[
\boxed{v(s) = 1 - \frac{f_Q}{\sigma_Q^2} \cdot (r_Q(s) - f_Q)}
\]
这就是 (8A-21) 的完整推导。