Module: OptionLab::BjerksundStensland
- Defined in:
- lib/option_lab/bjerksund_stensland.rb
Overview
Implementation of the Bjerksund-Stensland model for American options pricing Based on the 2002 improved version of their model
Class Method Summary collapse
-
.bjerksund_stensland_2002(s0, x, r, q, volatility, t1, t2) ⇒ Float
private
Core implementation of the Bjerksund-Stensland 2002 model.
-
.black_scholes_call(s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) ⇒ Float
private
Calculate the Black-Scholes price for a European call option.
-
.black_scholes_call_delta(s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) ⇒ Float
private
Calculate the Black-Scholes delta for a European call option.
-
.black_scholes_put(s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) ⇒ Float
private
Calculate the Black-Scholes price for a European put option.
-
.get_greeks(option_type, s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) ⇒ Hash
Calculate option Greeks using the Bjerksund-Stensland model and finite difference methods.
-
.max(a, b) ⇒ Float
private
Helper function to return maximum of two values.
-
.phi(s0, t, gamma, h, i, r, q, volatility) ⇒ Float
private
The phi function from the Bjerksund-Stensland model.
-
.price_american_call(s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) ⇒ Float
Price an American call option using the Bjerksund-Stensland model.
-
.price_american_put(s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) ⇒ Float
Price an American put option using the Bjerksund-Stensland model via put-call transformation.
-
.price_option(option_type, s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) ⇒ Float
Price an American option using the Bjerksund-Stensland model.
Class Method Details
.bjerksund_stensland_2002(s0, x, r, q, volatility, t1, t2) ⇒ Float (private)
Core implementation of the Bjerksund-Stensland 2002 model
159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 |
# File 'lib/option_lab/bjerksund_stensland.rb', line 159 def bjerksund_stensland_2002(s0, x, r, q, volatility, t1, t2) # Early exercise is never optimal if q <= 0 return black_scholes_call(s0, x, r, volatility, t2, q) if q <= 0 # To avoid domain errors with very small dividend yields return black_scholes_call(s0, x, r, volatility, t2, q) if q < 0.001 # Calculate parameters for the two-step approximation begin term1 = (r - q) / (volatility * volatility) term2 = (term1 - 0.5)**2 term3 = 2 * r / (volatility * volatility) beta = (0.5 - term1) + Math.sqrt(term2 + term3) b_inf = beta / (beta - 1) * x b_zero = max(x, r / q * x) # Calculate exercise boundaries for both time steps h1 = -(r - q) * t1 + 2 * volatility * Math.sqrt(t1) h2 = -(r - q) * t2 + 2 * volatility * Math.sqrt(t2) i1 = b_zero + (b_inf - b_zero) * (1 - Math.exp(h1)) i2 = b_zero + (b_inf - b_zero) * (1 - Math.exp(h2)) alpha1 = (i1 - x) * (i1**-beta) alpha2 = (i2 - x) * (i2**-beta) # Calculate the conditional risk-neutral probabilities result = if s0 >= i2 # Immediate exercise is optimal s0 - x elsif s0 >= i1 # Exercise at time t1 may be optimal alpha2 * (s0**beta) - alpha2 * phi(s0, t1, beta, i2, i2, r, q, volatility) + phi(s0, t1, 1, i2, i2, r, q, volatility) - phi(s0, t1, 1, x, i2, r, q, volatility) - x * phi(s0, t1, 0, i2, i2, r, q, volatility) + x * phi(s0, t1, 0, x, i2, r, q, volatility) + black_scholes_call(s0, x, r, volatility, t2, q) - black_scholes_call(s0, i2, r, volatility, t2, q) - (i2 - x) * black_scholes_call_delta(s0, i2, r, volatility, t2, q) else # Exercise at time t2 may be optimal alpha1 * (s0**beta) - alpha1 * phi(s0, t1, beta, i1, i2, r, q, volatility) + phi(s0, t1, 1, i1, i2, r, q, volatility) - phi(s0, t1, 1, x, i2, r, q, volatility) - x * phi(s0, t1, 0, i1, i2, r, q, volatility) + x * phi(s0, t1, 0, x, i2, r, q, volatility) + black_scholes_call(s0, x, r, volatility, t2, q) - black_scholes_call(s0, i2, r, volatility, t2, q) - (i2 - x) * black_scholes_call_delta(s0, i2, r, volatility, t2, q) end # Handle numerical issues - ensure result is not negative or NaN if !result.finite? || result < 0 # Fallback to Black-Scholes with a premium for early exercise bs_price = black_scholes_call(s0, x, r, volatility, t2, q) # Add a premium to represent the additional value of early exercise bs_price * (1.0 + q * t2 * 0.1) else result end rescue # Fallback to Black-Scholes with a premium for American features bs_price = black_scholes_call(s0, x, r, volatility, t2, q) # Add a premium to represent the additional value of early exercise bs_price * (1.0 + q * t2 * 0.1) end end |
.black_scholes_call(s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) ⇒ Float (private)
Calculate the Black-Scholes price for a European call option
233 234 235 236 237 238 239 240 241 242 243 |
# File 'lib/option_lab/bjerksund_stensland.rb', line 233 def black_scholes_call(s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) if years_to_maturity <= 0 return [s0 - x, 0.0].max end d1 = (Math.log(s0 / x) + (r - dividend_yield + 0.5 * volatility * volatility) * years_to_maturity) / (volatility * Math.sqrt(years_to_maturity)) d2 = d1 - volatility * Math.sqrt(years_to_maturity) s0 * Math.exp(-dividend_yield * years_to_maturity) * Distribution::Normal.cdf(d1) - x * Math.exp(-r * years_to_maturity) * Distribution::Normal.cdf(d2) end |
.black_scholes_call_delta(s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) ⇒ Float (private)
Calculate the Black-Scholes delta for a European call option
273 274 275 276 277 278 279 280 |
# File 'lib/option_lab/bjerksund_stensland.rb', line 273 def black_scholes_call_delta(s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) if years_to_maturity <= 0 return s0 >= x ? 1.0 : 0.0 end d1 = (Math.log(s0 / x) + (r - dividend_yield + 0.5 * volatility * volatility) * years_to_maturity) / (volatility * Math.sqrt(years_to_maturity)) Math.exp(-dividend_yield * years_to_maturity) * Distribution::Normal.cdf(d1) end |
.black_scholes_put(s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) ⇒ Float (private)
Calculate the Black-Scholes price for a European put option
253 254 255 256 257 258 259 260 261 262 263 |
# File 'lib/option_lab/bjerksund_stensland.rb', line 253 def black_scholes_put(s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) if years_to_maturity <= 0 return [x - s0, 0.0].max end d1 = (Math.log(s0 / x) + (r - dividend_yield + 0.5 * volatility * volatility) * years_to_maturity) / (volatility * Math.sqrt(years_to_maturity)) d2 = d1 - volatility * Math.sqrt(years_to_maturity) x * Math.exp(-r * years_to_maturity) * Distribution::Normal.cdf(-d2) - s0 * Math.exp(-dividend_yield * years_to_maturity) * Distribution::Normal.cdf(-d1) end |
.get_greeks(option_type, s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) ⇒ Hash
Calculate option Greeks using the Bjerksund-Stensland model and finite difference methods
133 134 135 136 137 138 139 140 141 142 143 144 145 146 |
# File 'lib/option_lab/bjerksund_stensland.rb', line 133 def get_greeks(option_type, s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) # Use the binomial tree model which is more reliable OptionLab::BinomialTree.get_greeks( option_type, s0, x, r, volatility, years_to_maturity, 100, # steps true, # American dividend_yield, ) end |
.max(a, b) ⇒ Float (private)
Helper function to return maximum of two values
309 310 311 |
# File 'lib/option_lab/bjerksund_stensland.rb', line 309 def max(a, b) a > b ? a : b end |
.phi(s0, t, gamma, h, i, r, q, volatility) ⇒ Float (private)
The phi function from the Bjerksund-Stensland model
292 293 294 295 296 297 298 299 300 301 302 303 |
# File 'lib/option_lab/bjerksund_stensland.rb', line 292 def phi(s0, t, gamma, h, i, r, q, volatility) lambda = (-r + gamma * (r - q) + 0.5 * gamma * (gamma - 1) * volatility * volatility) * t sqrt_t = Math.sqrt(t) d1 = -(Math.log(s0 / h) + (r - q + (gamma - 0.5) * volatility * volatility) * t) / (volatility * sqrt_t) d3 = -(Math.log(s0 / i) + (r - q + (gamma - 0.5) * volatility * volatility) * t) / (volatility * sqrt_t) s0**gamma * (Math.exp(lambda) * Distribution::Normal.cdf(-d1) - (i / h)**(2 * (r - q) / (volatility * volatility) - (2 * gamma - 1)) * Math.exp(lambda) * Distribution::Normal.cdf(-d3)) end |
.price_american_call(s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) ⇒ Float
Price an American call option using the Bjerksund-Stensland model
41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 |
# File 'lib/option_lab/bjerksund_stensland.rb', line 41 def price_american_call(s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) # If dividend yield is 0, American call = European call if dividend_yield <= 1e-10 return black_scholes_call(s0, x, r, volatility, years_to_maturity) end # If time to maturity is very small, return intrinsic value if years_to_maturity <= 1e-10 return [s0 - x, 0.0].max end # Use the 2002 improved version with two-step approximation # Split time to maturity in half for first step t1 = years_to_maturity / 2.0 t2 = years_to_maturity # Call the implementation with proper error handling begin result = bjerksund_stensland_2002(s0, x, r, dividend_yield, volatility, t1, t2) # Sanity check - ensure result is not negative if result < 0 # Fallback to Black-Scholes with a premium for early exercise bs_price = black_scholes_call(s0, x, r, volatility, years_to_maturity, dividend_yield) # Add a premium that increases with dividend yield and time to expiry result = bs_price * (1.0 + dividend_yield * years_to_maturity * 0.1) end result rescue # Fallback to Black-Scholes if there's a calculation error bs_price = black_scholes_call(s0, x, r, volatility, years_to_maturity, dividend_yield) # Add a premium that increases with dividend yield and time to expiry bs_price * (1.0 + dividend_yield * years_to_maturity * 0.1) end end |
.price_american_put(s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) ⇒ Float
Price an American put option using the Bjerksund-Stensland model via put-call transformation
84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 |
# File 'lib/option_lab/bjerksund_stensland.rb', line 84 def price_american_put(s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) # If time to maturity is very small, return intrinsic value if years_to_maturity <= 1e-10 return [x - s0, 0.0].max end # For simplicity, we'll use the binomial tree approach for American puts # which is more straightforward for put options begin result = OptionLab::BinomialTree.price_option( 'put', s0, x, r, volatility, years_to_maturity, 150, # Use a reasonable number of steps true, # It's an American option dividend_yield, ) # Sanity check - ensure the result is sensible if result < 0 || !result.finite? # Fallback to Black-Scholes with a premium for early exercise bs_price = black_scholes_put(s0, x, r, volatility, years_to_maturity, dividend_yield) # American put should always be more valuable than European put # Add a premium that increases with moneyness and time to expiry result = bs_price * (1.0 + 0.1 * years_to_maturity * (x > s0 ? (x - s0) / x : 0.01)) end result rescue # Fallback to Black-Scholes with a premium for early exercise bs_price = black_scholes_put(s0, x, r, volatility, years_to_maturity, dividend_yield) # American put should always be more valuable than European put # Add a premium that increases with moneyness and time to expiry bs_price * (1.0 + 0.1 * years_to_maturity * (x > s0 ? (x - s0) / x : 0.01)) end end |
.price_option(option_type, s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) ⇒ Float
Price an American option using the Bjerksund-Stensland model
22 23 24 25 26 27 28 29 30 31 |
# File 'lib/option_lab/bjerksund_stensland.rb', line 22 def price_option(option_type, s0, x, r, volatility, years_to_maturity, dividend_yield = 0.0) if option_type == 'call' price_american_call(s0, x, r, volatility, years_to_maturity, dividend_yield) elsif option_type == 'put' # Use put-call transformation for American puts price_american_put(s0, x, r, volatility, years_to_maturity, dividend_yield) else raise ArgumentError, "Option type must be either 'call' or 'put'!" end end |