关于「储备金破产精算:块状大修、重尾成本与有限期破产概率」,《分层公地与储备金精算》给出的定量结论是什么?
结论:第二部分,管理费原地不动。按一成计提:十年内见底,机会 0.1771;二十年内,0.6053;三十年内,高达 0.9454。即使计提率,恰好等于平均支出,也就是 0.1659,三十年内见底,机会仍有 0.3479。计提率提到 0.15,三十年见底机会 0.5172;提到 0.20,降到 0.1169;提到 0.25,只有 0.0250;提到 0.30,是 0.0107。
数据口径:大修支出块状、时点随机、偶有重尾意外;示意大楼即使管理费随通胀上调,按一成计提,三十年内见底机会仍有 0.5715,计提恰好等于平均支出,也有 0.3479。 书中楼盘、线索与参数全为合成或示意;全部计算由书中程序生成;本书是非个人化分析,不构成法律、税务或投资建议,法条与比例以上架前人工复核为准;线索只处理经同意者。
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: 0.0107, 0.0250, 0.1169, 0.1659, 0.1771, 0.3479
| mean spending, inflated, RM 238,860 a year | |||
| rate whose levy equals that mean: 0.1659 of charges | |||
| levy rate | ruin by yr 10 | by yr 20 | by yr 30 |
| 0.1000 | 0.1771 | 0.6053 | 0.9454 |
| 0.1659 | 0.0464 | 0.1297 | 0.3479 |
| 0.1500 | 0.0625 | 0.1966 | 0.5172 |
| 0.2000 | 0.0273 | 0.0503 | 0.1169 |
| 0.2500 | 0.0142 | 0.0176 | 0.0250 |
| 0.3000 | 0.0085 | 0.0094 | 0.0107 |
| charges raised with inflation; mean-covering rate 0.0935 | |||
| 0.1000 | 0.1358 | 0.3361 | 0.5715 |
| 0.0935 | 0.1591 | 0.4081 | 0.6724 |
| 0.1500 | 0.0460 | 0.0746 | 0.0954 |
| largest single year, median RM 1,812,910; 95th pct RM 4,041,391 |
核心公式
\[\psi(u) = \frac{e^{-Ru}}{1+\theta}, \quad R = \frac{\theta}{(1+\theta)\mu}.\]
查看完整验算程序
# Chapter 2. p22: finite-horizon ruin of a sinking fund, Monte Carlo.
# Same synthetic block as p21 (charges RM 1,440,000 a year). Spending is
# random in timing and in size:
# - each component is renewed after Weibull lifetimes with shape 3 and
# mean equal to its nominal cycle;
# - each renewal costs its base cost times a lognormal factor with
# median 1 and sigma 0.4 (overruns), inflated at g = 0.03 a year;
# - rare shocks (facade failure, flood, structural repair) arrive as a
# Poisson stream of 0.05 a year with Pareto sizes, alpha 1.8 and
# scale RM 100,000: a heavy tail.
# The fund earns r = 0.025. Ruin = a negative balance at any year end.
# A second table raises the charges, and so the levy, with inflation.
# Common random numbers: every levy rate meets the same 10,000 futures.
# Seed 3352. All parameters are illustrative.
import math
import random
CHARGES = 1440000.0
COMPONENTS = [(8, 312000.0), (15, 252000.0), (25, 960000.0),
(12, 156000.0), (15, 228000.0), (25, 270000.0),
(12, 180000.0)]
SHAPE, SIG, G, R = 3.0, 0.4, 0.03, 0.025
SHOCK_RATE, ALPHA, XM = 0.05, 1.8, 100000.0
YEARS, PATHS = 30, 10000
def future(rng):
spend = [[] for _ in range(YEARS + 1)]
for cyc, cost in COMPONENTS:
scale = cyc / math.gamma(1.0 + 1.0 / SHAPE)
t = rng.weibullvariate(scale, SHAPE)
while t <= YEARS:
yr = max(1, math.ceil(t))
spend[yr].append(cost * rng.lognormvariate(0.0, SIG))
t = t + rng.weibullvariate(scale, SHAPE)
t = rng.expovariate(SHOCK_RATE)
while t <= YEARS:
spend[max(1, math.ceil(t))].append(XM * rng.paretovariate(ALPHA))
t = t + rng.expovariate(SHOCK_RATE)
return [math.fsum(s) * (1.0 + G) ** k for k, s in enumerate(spend)]
def ruin_year(x, c, index=0.0):
b = 0.0
for t in range(1, YEARS + 1):
b = b * (1.0 + R) + c * (1.0 + index) ** t - x[t]
if b < 0.0:
return t
return None
if __name__ == "__main__":
rng = random.Random(3352)
futures = [future(rng) for _ in range(PATHS)]
yearly = math.fsum(math.fsum(x) for x in futures) / (PATHS * YEARS)
print("mean spending, inflated, RM {:,.0f} a year".format(yearly))
fair = yearly / CHARGES
print("rate whose levy equals that mean: {:.4f} of charges".format(
fair))
print("levy rate ruin by yr 10 by yr 20 by yr 30")
for rate in (0.10, fair, 0.15, 0.20, 0.25, 0.30):
hits = [ruin_year(x, rate * CHARGES) for x in futures]
row = []
for h in (10, 20, 30):
k = len([y for y in hits if y is not None and y <= h])
row.append(k / PATHS)
print("{:>9.4f} {:>13.4f} {:>8.4f} {:>8.4f}".format(
rate, row[0], row[1], row[2]))
real = math.fsum(math.fsum(v / (1.0 + G) ** k for k, v in
enumerate(x)) for x in futures)
fair2 = real / (PATHS * YEARS) / CHARGES
print("charges raised with inflation; mean-covering rate {:.4f}"
.format(fair2))
for rate in (0.10, fair2, 0.15):
hits = [ruin_year(x, rate * CHARGES, G) for x in futures]
row = []
for h in (10, 20, 30):
k = len([y for y in hits if y is not None and y <= h])
row.append(k / PATHS)
print("{:>9.4f} {:>13.4f} {:>8.4f} {:>8.4f}".format(
rate, row[0], row[1], row[2]))
big = [max(x) for x in futures]
big.sort()
print("largest single year, median RM {:,.0f}; 95th pct RM {:,.0f}"
.format(big[PATHS // 2], big[int(0.95 * PATHS)]))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《分层地契公地博弈、储备金破产精算与高意向置业线索派单母典》第 2 章「储备金破产精算:块状大修、重尾成本与有限期破产概率」(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-335 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026335,
author = {Fan, Yafu},
title = {Strata-Title Commons Games, Sinking-Fund Ruin Actuarial Science and High-Intent Property Lead Routing},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}