5D Library › 慢崩盘流动性储备、实物资产期权估值与跨代财富保全动力学 · English

第三章 困境资产的实物期权:要多大的折价才值得出手

等待本身有价值

买入一块土地、一座工厂或一家公司的机会,往往可以先留着、晚一点再用:这是一个投资的期权。McDonald 与 Siegel 以及 Dixit 与 Pindyck 的经典结果是:若资产价值服从几何布朗运动,只有当价值达到成本的 β/(β−1)\beta/(\beta-1) 倍时才值得出手 (McDonald and Siegel 1986; Dixit and Pindyck 1994)。换成困境收购的语言:若以资产价值减去折价 dd 的价格买入,折价至少要达到 1/β1/\beta。但困境报价不会永远等你:别的买家可能先出手。p05 把「报价每年被别人拿走的频率」加进去。

推导

# Chapter 3. p05: How deep a discount should a patient buyer demand? The
# chance to buy a real asset (land, a plant, a firm) can be held open
# and used later: an option to invest. If the asset's value V follows
# geometric Brownian motion with volatility sigma and payout (rent)
# yield q, and money earns r, buying is worth it only once
# V >= beta / (beta - 1) * price (McDonald and Siegel; Dixit and
# Pindyck). Paying price = (1 - d) V, the discount d must be at least
# 1 / beta. A distressed offer does not stay open for ever: if another
# buyer takes it at rate lam a year, beta solves
# sigma^2 b (b - 1) / 2 + (r - q) b - (r + lam) = 0.
# r = 0.05, q = 0.04; lam = 0 is the classic perpetual option.
from math import sqrt

def beta(sigma, lam=0.0, r=0.05, q=0.04):
    a = r - q - sigma * sigma / 2
    return (-a + sqrt(a * a + 2 * sigma * sigma * (r + lam))) / sigma ** 2

if __name__ == "__main__":
    lams = (0.0, 1.0, 4.0, 12.0)
    print("   discount needed before buying, by how often a year the")
    print("   offer is lost to another buyer")
    print("   volatility   lost 0   lost 1   lost 4   lost 12")
    for sigma in (0.15, 0.25, 0.35, 0.45):
        row = [1 / beta(sigma, lam) for lam in lams]
        print("   {:10.2f}   {:6.3f}   {:6.3f}   {:6.3f}   {:7.3f}".format(
            sigma, *row))
    b = beta(0.25)
    print("   perpetual, volatility 0.25: buy once V >= {:.2f} x price"
          .format(b / (b - 1)))
#    discount needed before buying, by how often a year the
#    offer is lost to another buyer
#    volatility   lost 0   lost 1   lost 4   lost 12
#          0.15    0.462    0.103    0.053     0.031
#          0.25    0.606    0.163    0.085     0.050
#          0.35    0.707    0.218    0.117     0.069
#          0.45    0.778    0.270    0.147     0.088
#    perpetual, volatility 0.25: buy once V >= 2.54 x price

如果报价永远不会被别人拿走,所需折价非常高:波动率 0.25 时是 0.606,资产价值要达到价格的 2.54 倍才出手。但只要报价每年有一次被拿走的机会,所需折价就降到 0.163;每年四次时只要 0.085。常说的「折价三成到五成才出手」,在这个模型里只对应很少有竞争买家的情况,或者波动率很高的情况:报价每年被拿走一次、波动率 0.45 时所需折价是 0.270。所需折价不是一个固定的数,而取决于你还有多少竞争者。

所需折价随报价被拿走的频率下降,灰色带为 0.3 至 0.5(p05)。

什么时候买,折价又值多少

p06 在第一章的路径上比较不同月份按市价买入、持有到第 60 个月的结果,以及在被迫卖出最集中的第 27 个月,以额外折价 dd 向被迫卖出者买入的结果。

推导

# Chapter 3. p06: When to buy, and what a forced seller's discount adds.
# On the 2,000 paths of p02 (seed 88), an asset is bought in month m at
# the market price and held to month 60. A distressed purchase in month
# 27 (between months 24 and 30, when p04's forced sales peak) is made
# at a further discount d to the market price. Reported: the median and
# 10th percentile of the value multiple at month 60.
import importlib.util
import random

def load(name):
    spec = importlib.util.spec_from_file_location(name, name + ".py")
    m = importlib.util.module_from_spec(spec)
    spec.loader.exec_module(m)
    return m

p02 = load("p02_slow_crash")

if __name__ == "__main__":
    rng = random.Random(88)
    paths = [p02.path(rng) for _ in range(2000)]
    q = lambda v, a: sorted(v)[int(a * (len(v) - 1))]
    print("   bought            median multiple   10th percentile")
    for m in (0, 6, 12, 18, 24, 30, 36):
        v = [p[60] / p[m] for p in paths]
        print("   month {:2d}          {:15.2f}   {:15.2f}".format(
            m, q(v, 0.5), q(v, 0.1)))
    for d in (0.3, 0.5):
        v = [p[60] / (p[27] * (1 - d)) for p in paths]
        print("   month 27, d {:.1f}   {:15.2f}   {:15.2f}".format(
            d, q(v, 0.5), q(v, 0.1)))
    v = [p[60] / min(p) for p in paths]
    print("   at the true low   {:15.2f}   {:15.2f}".format(
        q(v, 0.5), q(v, 0.1)))
#    bought            median multiple   10th percentile
#    month  0                     0.78              0.69
#    month  6                     0.82              0.68
#    month 12                     0.92              0.76
#    month 18                     1.06              0.88
#    month 24                     1.21              1.00
#    month 30                     1.39              1.14
#    month 36                     1.40              1.15
#    month 27, d 0.3              1.85              1.54
#    month 27, d 0.5              2.59              2.15
#    at the true low              1.58              1.34

第 0 个月买入,到第 60 个月的价值倍数中位数只有 0.78;第 24 个月买入是 1.21,第 30 个月是 1.39。即使事后知道真正的低点、在低点买入,中位数也只有 1.58。而在第 27 个月以额外折价三成向被迫卖出者买入,中位数是 1.85;额外折价五成是 2.59,最差的一成情况也有 2.15。对持有现金的买方,向被迫卖出者买到的折价,比猜中低点更值钱。

假说 · 尚未验证

作者的诠释(尚未验证): 在慢崩盘的后半段,向被迫卖出者私下协议收购的实物资产,其五年后的回报高于同期在公开市场买入的同类资产。什么会证明它错: 在困境处置的成交数据中,私下协议收购与公开市场买入同类资产的五年回报没有显著差别。

小结

实物期权的经典结果要求很深的折价,但竞争买家会让等待的价值迅速消失;「折价三成到五成」对应的是竞争很少的困境市场。真正的超额回报来自被迫卖出者给出的折价,而不是猜中低点。

参考文献

Dixit, Avinash K., and Robert S. Pindyck. 1994. Investment under Uncertainty. Princeton University Press.
McDonald, Robert, and Daniel Siegel. 1986. “The Value of Waiting to Invest.” The Quarterly Journal of Economics 101 (4): 707 to 727.

Fan, Y. (2026). Slow-Crash Liquidity Reserves, Real-Asset Option Valuation and Intergenerational Wealth Preservation: Thirty-Month Drawdown Kinetics, Distressed Resource Acquisition and Ergodic Dynasty Trusts [慢崩盘流动性储备、实物资产期权估值与跨代财富保全动力学] (Version 1.0.0). 5D Library. https://doi.org/10.5281/zenodo.24090231 · Licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0).