QuantLib 金融计算——案例之普通欧式期权分析
载入 QuantLib 和其他包:
import QuantLib as ql
import numpy as np
import pandas as pd
print(ql.__version__)
1.15
概述
从金融工程中最简单的案例——“普通欧式期权公式法定价”入手,介绍 QuantLib 中期权分析的基本组件,以及如何将这些组件拼接成为一个完整的计算流程。
普通欧式期权公式法定价
采用《期权、期货及其他衍生产品(第 7 版)》第 17 章中的例子:
- 类型:普通欧式看涨期权
- 当前价格:49$
- 敲定价格:50$
- 无风险利率:5%
- 年化波动率:20%
- 期限:20 周
使用 BS 公式为上述期权定价,并计算希腊值。
1. 配置期权合约条款
# 配置日期计算条款
calendar = ql.UnitedStates(ql.UnitedStates.NYSE)
dayCounter = ql.Actual365Fixed(ql.Actual365Fixed.Standard)
todayDate = ql.Date(11, ql.July, 2019)
maturity = todayDate + ql.Period(20, ql.Weeks)
settlementDate = todayDate
# 配置期权参数
stock = 49
strike = 50
riskFreeRate = 0.05
volatility = 0.2
# 配置全局估值日期
ql.Settings.instance().evaluationDate = todayDate
2. 构建期权对象
# 配置行权条款
europeanExercise = ql.EuropeanExercise(maturity)
optionType = ql.Option.Call
payoff = ql.PlainVanillaPayoff(
type=optionType, strike=strike)
# 构建期权对象
europeanOption = ql.VanillaOption(
payoff=payoff,
exercise=europeanExercise)
3. 配置定价引擎
underlying = ql.SimpleQuote(stock)
underlyingH = ql.QuoteHandle(underlying)
# 无风险利率的期限结构
flatRiskFreeTS = ql.YieldTermStructureHandle(
ql.FlatForward(
settlementDate, riskFreeRate, dayCounter))
# 波动率的期限结构
flatVolTS = ql.BlackVolTermStructureHandle(
ql.BlackConstantVol(
settlementDate, calendar,
volatility, dayCounter))
# 构建 BS 过程
bsProcess = ql.BlackScholesProcess(
s0=underlyingH,
riskFreeTS=flatRiskFreeTS,
volTS=flatVolTS)
# 基于 BS 过程的公式定价引擎
pricingEngine = ql.AnalyticEuropeanEngine(
bsProcess)
europeanOption.setPricingEngine(pricingEngine)
4. 计算
# RESULTS
print("Option value =", europeanOption.NPV())
print("Delta value =", europeanOption.delta())
print("Theta value =", europeanOption.theta())
print("Theta perday =", europeanOption.thetaPerDay())
print("Gamma value =", europeanOption.gamma())
print("Vega value =", europeanOption.vega())
print("Rho value =", europeanOption.rho())
Option value = 2.395988448539984
Delta value = 0.5213970624832108
Theta value = -4.309457134907618
Theta perday = -0.011806731876459226
Gamma value = 0.06563585494066533
Vega value = 12.089225358769994
Rho value = 8.88039853654583
题外话:天数计算规则
Quote
带来的便利
QuantLib 中有相当多的组件接受 Handle
类型的参数,而这些参数通常持有一个 Quote
类型的变量。借助“观察者模式”,用户修改 Quote
类型变量的值将会自动通知相关组件,并使其重新进行性计算,而无需再次构建一遍计算流程。对于某些用途来讲,这带来了相当大的便利。
# USE QUOTE
stock_array = np.arange(
start=30, stop=70, step=0.01)
NPV = np.array([np.nan] * len(stock_array))
delta = np.array([np.nan] * len(stock_array))
theta = np.array([np.nan] * len(stock_array))
# thetaPerDay = np.array([np.nan] * len(stock_array))
gamma = np.array([np.nan] * len(stock_array))
vega = np.array([np.nan] * len(stock_array))
rho = np.array([np.nan] * len(stock_array))
for i, v in enumerate(stock_array):
# 重置 Quote 对象的值
underlying.setValue(v)
# 无须再次配置计算流程,直接计算
NPV[i] = europeanOption.NPV()
delta[i] = europeanOption.delta()
theta[i] = europeanOption.theta()
# thetaPerDay[i] = europeanOption.thetaPerDay()
gamma[i] = europeanOption.gamma()
vega[i] = europeanOption.vega()
rho[i] = europeanOption.rho()
result = pd.DataFrame(
data=dict(
NPV=NPV,
delta=delta,
theta=theta,
# thetaPerDay=thetaPerDay,
gamma=gamma,
vega=vega, rho=rho),
index=stock_array)
result.plot(subplots=True)
总结
下面用一副图显示上述例子中的若干变量如何汇聚成一个计算流程: