关于「固定保费怎样花:行权水平、赔付上限、现金储备与参数不确定」,《尾部保护与凯利边界》给出的定量结论是什么?

结论:程序 p61 显示,同样花掉资本的 0.006,行权 0.70 的保护把期望短缺降到 0.2396、把资本跌破 0.7 的概率降到 0.00058,但增长率降到 0.13936;0.90 到 0.60 的封顶保护期望短缺只降到 0.2957,增长率却升到 0.14033,是唯一高于不买保护的方案。程序 p62 显示,把期望短缺压到 0.25,现金储备每年放弃 0.0370,行权 0.90 的保护只放弃 0.0033,两者的盈亏平衡点是经营回报比现金利率高 0.0136,但基差风险与流动性没有计入。程序 p63 显示,20 年里见过 0 次与 2 次跳跃的经营者,对同一报价会得出相反的结论,即使 100 年的记录,边界区间 0.01517 到 0.02524 仍包住报价。下一章回到古人,看《左传》的「有备无患」与亚里士多德笔下泰勒斯的榨油机,能被今天的模型检验出什么。

数据口径:用默顿跳扩散模型描述营收指数,并让保费报价对跳跃风险收取示意的溢价,行权 0.80 的固定保费保护,保费是平均赔付的 1.669 倍,买方平均收益为 -0.401;越只保极端情形,保费相对平均赔付越贵。 书中参数为示意;全部计算由书中程序生成;本书是非个人化法律、税务或理财建议,不推荐任何具体金融产品、保险产品、机构或交易策略,不承诺收益,不鼓励杠杆。

English version

对照数据表

下表数字由本章 Python 程序原样打印,可重新运行复现。 程序只用 Python 标准库与结论吻合的打印值: 0.00058, 0.006, 0.13936, 0.14033, 0.2396, 0.2957

budget 0.006 of capital, L = 1.0
shapehES 0.975VaR 0.99P(M<0.7)E log M
none0.0000.33730.33900.015800.14024
put 0.900.1470.29080.29230.009010.14024
put 0.800.2950.26730.26850.005530.13980
put 0.700.6590.23960.24020.000580.13936
spread 0.90/0.60 0.1600.29570.29690.009770.14033
spread 0.80/0.50 0.3100.26790.26470.005250.13985
查看完整验算程序
# Chapter 6. p61: one premium budget, five shapes of protection.
# The operator (L = 1.0) spends the same budget B = 0.006 of capital on
# either a plain protection struck at 0.90, 0.80 or 0.70, or a capped
# one (a spread: pays (K1 - S)+ - (K2 - S)+). Cover ratio h = B / (L *
# premium). Premiums at Q (p11); risk under P by Simpson quadrature.
# M = 1 + A + L * (S - 1) + h * L * (payout - premium). Illustrative.
import math

MU, SIG = 0.05, 0.18
LAM, JM, JS = 0.10, -0.35, 0.15
R, LAMQ = 0.03, 0.20
A, L, B = 0.12, 1.0, 0.006


def Phi(x):
    return 0.5 * math.erfc(-x / math.sqrt(2.0))


def comps(mu, lam):
    kap = math.exp(JM + 0.5 * JS * JS) - 1.0
    base = mu - 0.5 * SIG * SIG - lam * kap
    return [(math.exp(-lam) * lam ** n / math.factorial(n), base + n * JM,
             math.sqrt(SIG * SIG + n * JS * JS)) for n in range(13)]


def put_mix(cs, k):
    t = []
    for w, m, s in cs:
        d = (math.log(k) - m) / s
        t.append(w * (k * Phi(d) - math.exp(m + 0.5 * s * s) * Phi(d - s)))
    return math.fsum(t)


def nodes(cs, half=10.0, n=800):
    out = []
    for w, m, s in cs:
        h = 2.0 * half * s / n
        for i in range(n + 1):
            x = m - half * s + i * h
            c = 1 if i in (0, n) else (4 if i % 2 else 2)
            dens = math.exp(-0.5 * ((x - m) / s) ** 2) / (s * 2.5066282746)
            out.append((math.exp(x), w * c * h / 3.0 * dens))
    return sorted(out)


CP, CQ = comps(MU, LAM), comps(R, LAMQ)
ND = nodes(CP)


def s_quantile(a):
    lo, hi = 1e-6, 5.0
    for _ in range(100):
        mid = 0.5 * (lo + hi)
        c = math.fsum(w * Phi((math.log(mid) - m) / s) for w, m, s in CP)
        lo, hi = (mid, hi) if c < a else (lo, mid)
    return 0.5 * (lo + hi)


S01 = s_quantile(0.01)          # M 随 S 单调,VaR 由 S 的分位数精确给出


def run(k1, k2, budget):
    prem = math.exp(-R) * put_mix(CQ, k1)
    if k2:
        prem -= math.exp(-R) * put_mix(CQ, k2)
    h = budget / (L * prem)

    def cap(s):
        pay = max(k1 - s, 0.0) - (max(k2 - s, 0.0) if k2 else 0.0)
        return 1.0 + A + L * (s - 1.0) + h * L * (pay - prem)
    vals = sorted((cap(s), p) for s, p in ND)
    need, acc = 0.025, []
    for m, p in vals:
        use = min(p, need)
        acc.append(use * (1.0 - m))
        need -= use                                            # step
        if need <= 0.0:
            break
    es = math.fsum(acc) / 0.025
    deep = math.fsum(p for m, p in vals if m < 0.7)
    g = math.fsum(p * math.log(m) for m, p in vals)
    return h, es, 1.0 - cap(S01), deep, g


print("budget 0.006 of capital, L = 1.0")
print("shape            h      ES 0.975  VaR 0.99  P(M<0.7)  E log M")
for name, k1, k2, b in (("none", 0.9, None, 0.0),
                        ("put 0.90", 0.9, None, B),
                        ("put 0.80", 0.8, None, B),
                        ("put 0.70", 0.7, None, B),
                        ("spread 0.90/0.60", 0.9, 0.6, B),
                        ("spread 0.80/0.50", 0.8, 0.5, B)):
    h, es, v, d, g = run(k1, k2, b)
    print("{:16s} {:.3f}  {:.4f}    {:.4f}    {:.5f}   {:.5f}".format(
        name, h, es, v, d, g))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《尾部黑天鹅期权护城河与分位数对冲、跳扩散模型与凯利边界下的固定保费现金流保护母典》第 6 章「固定保费怎样花:行权水平、赔付上限、现金储备与参数不确定」(版本 1.0.0). 5D Library. DOI 待发布

DOI 待发布 · 专著页正式发布后上线(本站暂无专著页) · 5DL-B-290 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026290,
  author    = {Fan, Yafu},
  title     = {Tail-Risk Option Moats and Quantile Hedging: Jump-Diffusion Models and Fixed-Premium Protection of Core Cash Flows under Kelly Bounds},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论