Merge pull request #11 from PyFE/dev-0.1.3

Dev 0.1.3

Author: jaehyukchoi

.gitignore

@@ -1,4 +1,5 @@
 .DS_Store
+*_pyfe.py
 
 # Created by https://www.toptal.com/developers/gitignore/api/python,pycharm
 # Edit at https://www.toptal.com/developers/gitignore?templates=python,pycharm

README.md

@@ -29,7 +29,7 @@ pip install --upgrade pyfeng
 import numpy as np
 import pyfeng as pf
 m = pf.Bsm(sigma=0.2, intr=0.05, divr=0.1)
-m.price(np.arange(80, 121, 10), 100, 1.2)
+m.price(strike=np.arange(80, 121, 10), spot=100, texp=1.2)
 ```
 `Out [1]:`
 ```
@@ -38,15 +38,14 @@ array([15.71361973,  9.69250803,  5.52948546,  2.94558338,  1.48139131])
 
 `In [2]:`
 ```python
-sigma = np.array([0.2, 0.3, 0.5])[:, None]
+sigma = np.array([0.2], [0.5])
 m = pf.Bsm(sigma, intr=0.05, divr=0.1) # sigma in axis=0
-m.price(np.array([90, 100, 110]), 100, 1.2, cp=np.array([-1,1,1]))
+m.price(strike=np.array([90, 95, 100]), spot=100, texp=1.2, cp=np.array([-1,1,1]))
 ```
 `Out [2]:`
 ```
-array([[ 5.75927238,  5.52948546,  2.94558338],
-       [ 9.4592961 ,  9.3881245 ,  6.45745004],
-       [16.812035  , 17.10541288, 14.10354768]])
+array([[ 5.75927238,  7.38869609,  5.52948546],
+       [16.812035  , 18.83878533, 17.10541288]])
 ```
 
 ## Author

pyfeng/__init__.py

@@ -2,4 +2,6 @@
 from .bsm import Bsm, BsmDisp
 from .cev import Cev
 from .sabr import SabrHagan2002, SabrLorig2017, SabrChoiWu2021H, SabrChoiWu2021P
+from .sabr_int import SabrUncorrChoiWu2021
 from .nsvh import Nsvh1
+from .multiasset import BsmSpreadKirk, BsmSpreadBjerksund2014, NormBasket, NormSpread, BsmBasketLevy1992, BsmMax2

pyfeng/cev.py

@@ -5,7 +5,7 @@
 from . import opt_smile_abc as smile
 
 
-class Cev(opt.OptAnalyticABC, smile.OptSmileABC):
+class Cev(opt.OptAnalyticABC, smile.OptSmileABC, smile.MassZeroABC):
     """
     Constant Elasticity of Variance (CEV) model.
 
@@ -41,18 +41,7 @@ def params_kw(self):
         return {**params, **extra}  # Py 3.9, params | extra
 
     def mass_zero(self, spot, texp, log=False):
-        """
-        Probability mass at zero (absorbing boundary)
-
-        Args:
-            spot: spot or forward
-            texp: time to expiry
-            log: log value if True
-
-        Returns:
-            mass at zero.
-        """
-        fwd, _, _ = self._fwd_factor(spot, texp)
+        fwd = self.forward(spot, texp)
 
         betac = 1.0 - self.beta
         a = 0.5 / betac
@@ -68,6 +57,22 @@ def mass_zero(self, spot, texp, log=False):
         else:
             return spst.gamma.sf(x=x, a=a)
 
+    def mass_zero_t0(self, spot, texp):
+        """
+        Limit value of -T log(M_T) as T -> 0, where M_T is the mass at zero.
+
+        Args:
+            spot: spot (or forward) price
+
+        Returns:
+            - lim_{T->0} T log(M_T)
+        """
+        fwd = self.forward(spot, texp)
+        betac = 1.0 - self.beta
+        alpha = self.sigma/np.power(fwd, betac)
+        t0 = 0.5/(betac*alpha)**2
+        return t0
+
     @staticmethod
     def price_formula(strike, spot, texp, cp=1, sigma=None, beta=0.5, intr=0.0, divr=0.0, is_fwd=False):
         """

pyfeng/multiasset.py

@@ -0,0 +1,248 @@
+import scipy.stats as scst
+import numpy as np
+from . import bsm
+from . import norm
+from . import opt_abc as opt
+
+
+class BsmSpreadKirk(opt.OptMaABC):
+    """
+    Kirk's approximation for spread option.
+
+    References:
+        Kirk, E. (1995). Correlation in the energy markets. In Managing Energy Price Risk
+        (First, pp. 71–78). Risk Publications.
+
+    Examples:
+        >>> import numpy as np
+        >>> import pyfeng as pf
+        >>> m = pf.BsmSpreadKirk((0.2, 0.3), cor=-0.5)
+        >>> m.price(np.arange(-2, 3) * 10, [100, 120], 1.3)
+        array([22.15632247, 17.18441817, 12.98974214,  9.64141666,  6.99942072])
+    """
+
+    weight = np.array([1, -1])
+
+    def price(self, strike, spot, texp, cp=1):
+        df = np.exp(-texp * self.intr)
+        fwd = np.array(spot) * (1.0 if self.is_fwd else np.exp(-texp * np.array(self.divr)) / df)
+        assert fwd.shape[-1] == self.n_asset
+
+        fwd1 = fwd[..., 0] - np.minimum(strike, 0)
+        fwd2 = fwd[..., 1] + np.maximum(strike, 0)
+
+        sig1 = self.sigma[0] * fwd[..., 0] / fwd1
+        sig2 = self.sigma[1] * fwd[..., 1] / fwd2
+        sig_spd = np.sqrt(sig1*(sig1 - 2.0*self.rho*sig2) + sig2**2)
+        price = bsm.Bsm.price_formula(fwd2, fwd1, sig_spd, texp, cp=cp, is_fwd=True)
+        return df * price
+
+
+class BsmSpreadBjerksund2014(opt.OptMaABC):
+    """
+    Bjerksund & Stensland (2014)'s approximation for spread option.
+
+    References:
+        Bjerksund, P., & Stensland, G. (2014). Closed form spread option valuation.
+        Quantitative Finance, 14(10), 1785–1794. https://doi.org/10.1080/14697688.2011.617775
+
+    Examples:
+        >>> import numpy as np
+        >>> import pyfeng as pf
+        >>> m = pf.BsmSpreadBjerksund2014((0.2, 0.3), cor=-0.5)
+        >>> m.price(np.arange(-2, 3) * 10, [100, 120], 1.3)
+        array([22.13172022, 17.18304247, 12.98974214,  9.54431944,  6.80612597])
+    """
+
+    weight = np.array([1, -1])
+
+    def price(self, strike, spot, texp, cp=1):
+        df = np.exp(-texp * self.intr)
+        fwd = np.array(spot) * (1.0 if self.is_fwd else np.exp(-texp * np.array(self.divr)) / df)
+        assert fwd.shape[-1] == self.n_asset
+
+        fwd1 = fwd[..., 0]
+        fwd2 = fwd[..., 1]
+
+        std11 = self.sigma[0]**2 * texp
+        std12 = self.sigma[0]*self.sigma[1] * texp
+        std22 = self.sigma[1]**2 * texp
+
+        aa = fwd2 + strike
+        bb = fwd2/aa
+        std = np.sqrt(std11 - 2*bb*self.rho*std12 + bb**2*std22)
+
+        d3 = np.log(fwd1/aa)
+        d1 = (d3 + 0.5*std11 - bb*(self.rho*std12 - 0.5*bb*std22)) / std
+        d2 = (d3 - 0.5*std11 + self.rho*std12 + bb*(0.5*bb - 1)*std22) / std
+        d3 = (d3 - 0.5*std11 + 0.5*bb**2*std22) / std
+
+        price = cp*(fwd1*scst.norm.cdf(cp*d1) - fwd2*scst.norm.cdf(cp*d2)
+                    - strike*scst.norm.cdf(cp*d3))
+
+        return df * price
+
+
+class NormBasket(opt.OptMaABC):
+    """
+    Basket option pricing under the multiasset Bachelier model
+    """
+
+    weight = None
+
+    def __init__(self, sigma, cor=None, weight=None, intr=0.0, divr=0.0, is_fwd=False):
+        """
+        Args:
+            sigma: model volatilities of `n_asset` assets. (n_asset, ) array
+            cor: correlation. If matrix, used as it is. (n_asset, n_asset)
+                If scalar, correlation matrix is constructed with all same off-diagonal values.
+            weight: asset weights, If None, equally weighted as 1/n_asset
+                If scalar, equal weights of the value
+                If 1-D array, uses as it is. (n_asset, )
+            intr: interest rate (domestic interest rate)
+            divr: vector of dividend/convenience yield (foreign interest rate) 0-D or (n_asset, ) array
+            is_fwd: if True, treat `spot` as forward price. False by default.
+        """
+
+        super().__init__(sigma, cor=cor, intr=intr, divr=divr, is_fwd=is_fwd)
+        if weight is None:
+            self.weight = np.ones(self.n_asset) / self.n_asset
+        elif np.isscalar(weight):
+            self.weight = np.ones(self.n_asset) * weight
+        else:
+            assert len(weight) == self.n_asset
+            self.weight = np.array(weight)
+
+    def price(self, strike, spot, texp, cp=1):
+        df = np.exp(-texp * self.intr)
+        fwd = np.array(spot) * (1.0 if self.is_fwd else np.exp(-texp * np.array(self.divr)) / df)
+        assert fwd.shape[-1] == self.n_asset
+
+        fwd_basket = fwd @ self.weight
+        vol_basket = np.sqrt(self.weight @ self.cov_m @ self.weight)
+
+        price = norm.Norm.price_formula(
+            strike, fwd_basket, vol_basket, texp, cp=cp, is_fwd=True)
+        return df * price
+
+
+class NormSpread(opt.OptMaABC):
+    """
+    Spread option pricing under the Bachelier model.
+    This is a special case of NormBasket with weight = (1, -1)
+
+    Examples:
+        >>> import numpy as np
+        >>> import pyfeng as pf
+        >>> m = pf.NormSpread((20, 30), cor=-0.5, intr=0.05)
+        >>> m.price(np.arange(-2, 3) * 10, [100, 120], 1.3)
+        array([17.95676186, 13.74646821, 10.26669936,  7.47098719,  5.29057157])
+    """
+    weight = np.array([1, -1])
+
+    price = NormBasket.price
+
+
+class BsmBasketLevy1992(NormBasket):
+    """
+    Basket option pricing with the log-normal approximation of Levy & Turnbull (1992)
+
+    References:
+        Levy, E., & Turnbull, S. (1992). Average intelligence. Risk, 1992(2), 53–57.
+
+        Krekel, M., de Kock, J., Korn, R., & Man, T.-K. (2004). An analysis of pricing methods for basket options.
+        Wilmott Magazine, 2004(7), 82–89.
+
+    Examples:
+        >>> import numpy as np
+        >>> import pyfeng as pf
+        >>> strike = np.arange(50, 151, 10)
+        >>> m = pf.BsmBasketLevy1992(sigma=0.4*np.ones(4), cor=0.5)
+        >>> m.price(strike, spot=100*np.ones(4), texp=5)
+        array([54.34281026, 47.521086  , 41.56701301, 36.3982413 , 31.92312156,
+               28.05196621, 24.70229571, 21.800801  , 19.28360474, 17.09570196,
+               15.19005654])
+    """
+    def price(self, strike, spot, texp, cp=1):
+        df = np.exp(-texp * self.intr)
+        fwd = np.array(spot) * (1.0 if self.is_fwd else np.exp(-texp * np.array(self.divr)) / df)
+        assert fwd.shape[-1] == self.n_asset
+
+        fwd_basket = fwd * self.weight
+        m1 = np.sum(fwd_basket, axis=-1)
+        m2 = np.sum(fwd_basket @ np.exp(self.cov_m * texp) * fwd_basket, axis=-1)
+
+        sig = np.sqrt(np.log(m2/(m1**2))/texp)
+        price = bsm.Bsm.price_formula(
+            strike, m1, sig, texp, cp=cp, is_fwd=True)
+        return df * price
+
+
+class BsmMax2(opt.OptMaABC):
+    """
+    Option on the max of two assets.
+    Payout = max( max(F_1, F_2) - K, 0 ) for all or  max( K - max(F_1, F_2), 0 ) for put option
+
+    References:
+        Rubinstein, M. (1991). Somewhere Over the Rainbow. Risk, 1991(11), 63–66.
+
+    Examples:
+        >>> import numpy as np
+        >>> import pyfeng as pf
+        >>> m = pf.BsmMax2(0.2*np.ones(2), cor=0, divr=0.1, intr=0.05)
+        >>> m.price(strike=[90, 100, 110], spot=100*np.ones(2), texp=3)
+        array([15.86717049, 11.19568103,  7.71592217])
+    """
+
+    m_switch = None
+
+    def __init__(self, sigma, cor=None, weight=None, intr=0.0, divr=0.0, is_fwd=False):
+        super().__init__(sigma, cor=cor, intr=intr, divr=divr, is_fwd=is_fwd)
+        self.m_switch = BsmSpreadKirk(sigma, cor, is_fwd=True)
+
+    def price(self, strike, spot, texp, cp=1):
+        sig = self.sigma
+        df = np.exp(-texp * self.intr)
+        fwd = np.array(spot) * (1.0 if self.is_fwd else np.exp(-texp * np.array(self.divr)) / df)
+        assert fwd.shape[-1] == self.n_asset
+
+        sig_std = sig * np.sqrt(texp)
+        spd_rho = np.sqrt(np.dot(sig, sig) - 2*self.rho*sig[0]*sig[1])
+        spd_std = spd_rho * np.sqrt(texp)
+
+        # -x and y as rows
+        # supposed to be -log(fwd/strike) but strike is added later
+        xx = -np.log(fwd)/sig_std - 0.5*sig_std
+
+        fwd_ratio = fwd[0]/fwd[1]
+        yy = np.log([fwd_ratio, 1/fwd_ratio])/spd_std + 0.5*spd_std
+
+        rho12 = np.array([self.rho*sig[1]-sig[0], self.rho*sig[0]-sig[1]])/spd_rho
+
+        mu0 = np.zeros(2)
+        cor_m1 = rho12[0] + (1-rho12[0])*np.eye(2)
+        cor_m2 = rho12[1] + (1-rho12[1])*np.eye(2)
+
+        strike_isscalar = np.isscalar(strike)
+        strike = np.atleast_1d(strike)
+        cp = cp * np.ones_like(strike)
+        n_strike = len(strike)
+
+        # this is the price of max(S1, S2) = max(S1-S2, 0) + S2
+        # Used that Kirk approximation strike = 0 is Margrabe's switch option price
+        parity = 0 if np.all(cp > 0) else self.m_switch.price(0, fwd, texp) + fwd[1]
+        price = np.zeros_like(strike, float)
+        for k in range(n_strike):
+            xx_ = xx + np.log(strike[k])/sig_std
+            term1 = fwd[0] * (scst.norm.cdf(yy[0]) - scst.multivariate_normal.cdf(np.array([xx_[0], yy[0]]), mu0, cor_m1))
+            term2 = fwd[1] * (scst.norm.cdf(yy[1]) - scst.multivariate_normal.cdf(np.array([xx_[1], yy[1]]), mu0, cor_m2))
+            term3 = strike[k] * np.array(scst.multivariate_normal.cdf(xx_ + sig_std, mu0, self.cor_m))
+
+            assert(term1 + term2 + term3 >= strike[k])
+            price[k] = (term1 + term2 + term3 - strike[k])
+
+            if cp[k] < 0:
+                price[k] += strike[k] - parity
+        price *= df
+
+        return price[0] if strike_isscalar else price