Steven Shreve: Stochastic Calculus and Finance

P RASAD C HALASANI Carnegie Mellon University chal@cs.cmu.edu S OMESH J HA Carnegie Mellon University sjha@cs.cmu.edu

THIS IS A DRAFT: PLEASE DO NOT DISTRIBUTE c Copyright; Steven E. Shreve, 1996 October 6, 1997

Contents
1 Introduction to Probability Theory 1.1 1.2 1.3 1.4 1.5 The Binomial Asset Pricing Model . . . . . . . . . . . . . . . . . . . . . . . . . . Finite Probability Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lebesgue Measure and the Lebesgue Integral . . . . . . . . . . . . . . . . . . . . General Probability Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Independence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.5.1 1.5.2 1.5.3 1.5.4 1.5.5 1.5.6 1.5.7 Independence of sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Independence of -algebras . . . . . . . . . . . . . . . . . . . . . . . . . Independence of random variables . . . . . . . . . . . . . . . . . . . . . . Correlation and independence . . . . . . . . . . . . . . . . . . . . . . . . Independence and conditional expectation. . . . . . . . . . . . . . . . . . Law of Large Numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . Central Limit Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 11 16 22 30 40 40 41 42 44 45 46 47 49 49 50 52 52 53 54 55 57 58

2 Conditional Expectation 2.1 2.2 2.3 A Binomial Model for Stock Price Dynamics . . . . . . . . . . . . . . . . . . . . Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conditional Expectation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.1 2.3.2 2.3.3 2.3.4 2.3.5 2.4 An example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Definition of Conditional Expectation . . . . . . . . . . . . . . . . . . . . Further discussion of Partial Averaging . . . . . . . . . . . . . . . . . . . Properties of Conditional Expectation . . . . . . . . . . . . . . . . . . . . Examples from the Binomial Model . . . . . . . . . . . . . . . . . . . . .

Martingales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

2 3 Arbitrage Pricing 3.1 3.2 3.3 Binomial Pricing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . General one-step APT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Risk-Neutral Probability Measure . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1 3.3.2 3.4 3.5 Portfolio Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Self-financing Value of a Portfolio Process . . . . . . . . . . . . . . . . 59 59 60 61 62 62 63 64 67 67 69 70 70 73 74 77 77 79 81 85 85 86 88 89 91 91 92 94 97 97

Simple European Derivative Securities . . . . . . . . . . . . . . . . . . . . . . . . The Binomial Model is Complete . . . . . . . . . . . . . . . . . . . . . . . . . . .

4 The Markov Property 4.1 4.2 4.3 4.4 4.5 Binomial Model Pricing and Hedging . . . . . . . . . . . . . . . . . . . . . . . . Computational Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Markov Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.1 Different ways to write the Markov property . . . . . . . . . . . . . . . . Showing that a process is Markov . . . . . . . . . . . . . . . . . . . . . . . . . . Application to Exotic Options . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5 Stopping Times and American Options 5.1 5.2 5.3 American Pricing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Value of Portfolio Hedging an American Option . . . . . . . . . . . . . . . . . . . Information up to a Stopping Time . . . . . . . . . . . . . . . . . . . . . . . . . .

6 Properties of American Derivative Securities 6.1 6.2 6.3 6.4 The properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Proofs of the Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Compound European Derivative Securities . . . . . . . . . . . . . . . . . . . . . . Optimal Exercise of American Derivative Security . . . . . . . . . . . . . . . . . .

7 Jensen’s Inequality 7.1 7.2 7.3 Jensen’s Inequality for Conditional Expectations . . . . . . . . . . . . . . . . . . . Optimal Exercise of an American Call . . . . . . . . . . . . . . . . . . . . . . . . Stopped Martingales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

8 Random Walks 8.1 First Passage Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3 8.2 8.3 8.4 8.5 8.6 8.7 8.8 8.9 is almost surely finite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The moment generating function for Expectation of . . . . . . . . . . . . . . . . . . . . . . . . 97 99

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

The Strong Markov Property . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 General First Passage Times . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 Example: Perpetual American Put . . . . . . . . . . . . . . . . . . . . . . . . . . 102 Difference Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 Distribution of First Passage Times . . . . . . . . . . . . . . . . . . . . . . . . . . 107

8.10 The Reflection Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 9 Pricing in terms of Market Probabilities: The Radon-Nikodym Theorem. 9.1 9.2 9.3 9.4 9.5 111

Radon-Nikodym Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Radon-Nikodym Martingales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 The State Price Density Process . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Stochastic Volatility Binomial Model . . . . . . . . . . . . . . . . . . . . . . . . . 116 Another Applicaton of the Radon-Nikodym Theorem . . . . . . . . . . . . . . . . 118 119

10 Capital Asset Pricing

10.1 An Optimization Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 11 General Random Variables 123

11.1 Law of a Random Variable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 11.2 Density of a Random Variable . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 11.3 Expectation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 11.4 Two random variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 11.5 Marginal Density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 11.6 Conditional Expectation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 11.7 Conditional Density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 11.8 Multivariate Normal Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 11.9 Bivariate normal distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 11.10MGF of jointly normal random variables . . . . . . . . . . . . . . . . . . . . . . . 130 12 Semi-Continuous Models 131

12.1 Discrete-time Brownian Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

4 12.2 The Stock Price Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 12.3 Remainder of the Market . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 12.4 Risk-Neutral Measure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 12.5 Risk-Neutral Pricing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 12.6 Arbitrage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 12.7 Stalking the Risk-Neutral Measure . . . . . . . . . . . . . . . . . . . . . . . . . . 135 12.8 Pricing a European Call . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 13 Brownian Motion 139

13.1 Symmetric Random Walk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 13.2 The Law of Large Numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 13.3 Central Limit Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 13.4 Brownian Motion as a Limit of Random Walks . . . . . . . . . . . . . . . . . . . 141 13.5 Brownian Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 13.6 Covariance of Brownian Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 13.7 Finite-Dimensional Distributions of Brownian Motion . . . . . . . . . . . . . . . . 144 13.8 Filtration generated by a Brownian Motion . . . . . . . . . . . . . . . . . . . . . . 144 13.9 Martingale Property . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 13.10The Limit of a Binomial Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 13.11Starting at Points Other Than 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 13.12Markov Property for Brownian Motion . . . . . . . . . . . . . . . . . . . . . . . . 147 13.13Transition Density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 13.14First Passage Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 14 The Itˆ Integral o 153

14.1 Brownian Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 14.2 First Variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 14.3 Quadratic Variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 14.4 Quadratic Variation as Absolute Volatility . . . . . . . . . . . . . . . . . . . . . . 157 14.5 Construction of the Itˆ Integral . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 o 14.6 Itˆ integral of an elementary integrand . . . . . . . . . . . . . . . . . . . . . . . . 158 o 14.7 Properties of the Itˆ integral of an elementary process . . . . . . . . . . . . . . . . 159 o 14.8 Itˆ integral of a general integrand . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 o

5 14.9 Properties of the (general) Itˆ integral . . . . . . . . . . . . . . . . . . . . . . . . 163 o 14.10Quadratic variation of an Itˆ integral . . . . . . . . . . . . . . . . . . . . . . . . . 165 o 15 Itˆ ’s Formula o 167

15.1 Itˆ ’s formula for one Brownian motion . . . . . . . . . . . . . . . . . . . . . . . . 167 o 15.2 Derivation of Itˆ ’s formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 o 15.3 Geometric Brownian motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 15.4 Quadratic variation of geometric Brownian motion . . . . . . . . . . . . . . . . . 170 15.5 Volatility of Geometric Brownian motion . . . . . . . . . . . . . . . . . . . . . . 170 15.6 First derivation of the Black-Scholes formula . . . . . . . . . . . . . . . . . . . . 170 15.7 Mean and variance of the Cox-Ingersoll-Ross process . . . . . . . . . . . . . . . . 172 15.8 Multidimensional Brownian Motion . . . . . . . . . . . . . . . . . . . . . . . . . 173 15.9 Cross-variations of Brownian motions . . . . . . . . . . . . . . . . . . . . . . . . 174 15.10Multi-dimensional Itˆ formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 o 16 Markov processes and the Kolmogorov equations 177

16.1 Stochastic Differential Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 16.2 Markov Property . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178 16.3 Transition density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 16.4 The Kolmogorov Backward Equation . . . . . . . . . . . . . . . . . . . . . . . . 180 16.5 Connection between stochastic calculus and KBE . . . . . . . . . . . . . . . . . . 181 16.6 Black-Scholes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183 16.7 Black-Scholes with price-dependent volatility . . . . . . . . . . . . . . . . . . . . 186 17 Girsanov’s theorem and the risk-neutral measure 17.1 Conditional expectations under

f IP

189

. . . . . . . . . . . . . . . . . . . . . . . . . . 191

17.2 Risk-neutral measure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 18 Martingale Representation Theorem 197

18.1 Martingale Representation Theorem . . . . . . . . . . . . . . . . . . . . . . . . . 197 18.2 A hedging application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 18.3 18.4

d-dimensional Girsanov Theorem . . . . . . . . . . . d-dimensional Martingale Representation Theorem . .

. . . . . . . . . . . . . . . 199 . . . . . . . . . . . . . . . 200

18.5 Multi-dimensional market model . . . . . . . . . . . . . . . . . . . . . . . . . . . 200

6 19 A two-dimensional market model 19.1 Hedging when ;1 < 19.2 Hedging when 203 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204

=1

<1 .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205 209

20 Pricing Exotic Options

20.1 Reflection principle for Brownian motion . . . . . . . . . . . . . . . . . . . . . . 209 20.2 Up and out European call. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212 20.3 A practical issue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218 21 Asian Options 219

21.1 Feynman-Kac Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220 21.2 Constructing the hedge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220 21.3 Partial average payoff Asian option . . . . . . . . . . . . . . . . . . . . . . . . . . 221 22 Summary of Arbitrage Pricing Theory 223

22.1 Binomial model, Hedging Portfolio . . . . . . . . . . . . . . . . . . . . . . . . . 223 22.2 Setting up the continuous model . . . . . . . . . . . . . . . . . . . . . . . . . . . 225 22.3 Risk-neutral pricing and hedging . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 22.4 Implementation of risk-neutral pricing and hedging . . . . . . . . . . . . . . . . . 229 23 Recognizing a Brownian Motion 233

23.1 Identifying volatility and correlation . . . . . . . . . . . . . . . . . . . . . . . . . 235 23.2 Reversing the process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236 24 An outside barrier option 239

24.1 Computing the option value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242 24.2 The PDE for the outside barrier option . . . . . . . . . . . . . . . . . . . . . . . . 243 24.3 The hedge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245 25 American Options 247

25.1 Preview of perpetual American put . . . . . . . . . . . . . . . . . . . . . . . . . . 247 25.2 First passage times for Brownian motion: first method . . . . . . . . . . . . . . . . 247 25.3 Drift adjustment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249 25.4 Drift-adjusted Laplace transform . . . . . . . . . . . . . . . . . . . . . . . . . . . 250 25.5 First passage times: Second method . . . . . . . . . . . . . . . . . . . . . . . . . 251

. . . . . . . . . . . . .5 Forward-future spread . . . . . . . . . . . . . . . . . . . .5 Recovering the interest r(t) from the forward rate . . . . . . . forward contracts and futures 27. . . . . . . . 279 28. . 280 28. . . . . . . .2 Some interest-rate dependent assets . . .1 Forward contracts . . .8 Implementation of the Heath-Jarrow-Morton model . . . . .6 Backwardation and contango . . . . . . . . . 277 28. .6 Perpetual American put . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272 27. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11Put with expiration . . . . . . . . . . . . . . . . . . .1 An example: Brownian Motion . 261 26 Options on dividend-paying stocks 263 26. . . . . . . . . . . . . . 259 25. . 252 25. . . . . .4 Forward rate agreement . . . . . . . . . . . . . . . . . .3 Hedging at time t1 . . . . . . . .3 Future contracts . . .1 Computing arbitrage-free bond prices: first method . . . . .3 Terminology . . 257 25. . .7 25. . . . . . 269 27. . . . . . . . . . . 266 267 27 Bonds. . . . . . . . . .8 Hedging the put . . . . . . . . . . . . . . . . . . . 273 28 Term-structure models 275 28. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281 29 Gaussian processes 285 29. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 Checking for absence of arbitrage . . 270 27. . . . .12American contingent claim with expiration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269 27. . . . . . . . . . . . . . . . . . . . . . 259 25. .7 Value of the perpetual American put . . . . . . . . . . . . . .6 Computing arbitrage-free bond prices: Heath-Jarrow-Morton method . . . . . . . 260 25. . . .10Perpetual American call . . . . . . . 276 28. . . . . . . . . . . . . . . . . 278 28. . . . . . . . . . . . .4 Cash flow from a future contract . . 276 28. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263 26. . . . . . . . . . . . . . 272 27. . 264 26. . . . . . . . . . . . . . . . . . . . . 277 28. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286 30 Hull and White model 293 . . .2 Dividend paying stock . . . . . . . . . . . . . . . . 256 25. . .1 American option with convex payoff function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2 Hedging a forward contract . . . . . . . . . . . . . . . . . . . . . . . . .9 Perpetual American contingent claim . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5 Option on a bond . . . . . . . . . . . . . . . . . . . . 338 34. .2 Kolmogorov forward equation . . . . . . . . . . . . . . . . . . . . 313 31. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 Option on a bond . . . . . . . . . . . . . . . . . 306 31. . 340 34. . . . . . . . . . . . 310 31. . . . . . . . . . . . . . . . . . . . . . . . . . 343 .2 Zero-coupon bond prices . . . 309 31. . . . . .8 Tracking down '0(0) in the time change of the CIR model . . .7 Calibration . . . . . .2 Dynamics of the bond price . . . . . . . . . . . . 297 30. . . 336 34. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328 e 33. . . . . . . . 321 32. . . . . . . . . . . . . . . 338 34. . . . . . . . . . . . . 342 34. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315 31. . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 Bond prices . . . . . . .1 Bond price as num´ raire . . . . . . . . . . . . .5 The dynamics of L(t ) . . . . . . . . .3 Calibration .1 Review of HJM under risk-neutral IP . . . . . .6 Deterministic time change of CIR model . . . . . 295 30. . . . 316 32 A two-factor model (Duffie & Kan) 319 32. . . . . . 323 33 Change of num´ raire e 325 33. . . .4 Bond prices in the CIR model . . . . . . . 299 31 Cox-Ingersoll-Ross model 303 31. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335 34. . . . . .1 Equilibrium distribution of r(t) . . . . . . . . . .3 Calibration of the Hull & White model . 306 31. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327 e 33. . .3 Cox-Ingersoll-Ross equilibrium density . . . . . . . . 296 30. . . . . . . . . . . .8 Forward LIBOR under more forward measure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Fiddling with the formulas . 337 34. 313 31. . . . . . . . . . . 329 34 Brace-Gatarek-Musiela model 335 34. . . . . . . . . . . . 320 32. . . .8 30. . . . . . . . . . . . . . . . . . . . . . . . . . . .3 LIBOR . . . . . . . . . . . . . . . . . . . .2 Brace-Gatarek-Musiela model . . . . . . . . .4 Forward LIBOR . . . . . . . . . . . .1 Non-negativity of Y . . .6 Implementation of BGM . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2 Stock price as num´ raire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 Merton option pricing formula . . . . . .

.9 American options. .12Long rates . . . . . . . . . . . . . . . . . 349 35. . . 352 35. . . . . . . . . . . . . . the martingale representation theorem. . . . . 351 35. . . . . . . . . . . . .351 35. . . . . . . 353 35. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Probability theory and martingales. . . . .4 Stochastic integrals. . . . . . . . . . . . . .10Pricing an interest rate cap . . . .8 Exotic options. .9 34. . . . . . . . . . . . . 350 35. . . . . . . . . . . . . . . . . . . . . . . . . . 349 35. . . . 352 35. . . . . . . . . . . . . . . . . . . . . . . . . and risk-neutral measures. . . . . . . . . . . . . . . .13Pricing a swap . . . . . . . 346 34. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11Term structure models.14REFERENCES . .10Forward and futures contracts. . . . . . . . . . . . . . . . . 354 . . . . . . . . . . . . . . . . . .2 Binomial asset pricing model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 Girsanov’s theorem. . . . . . . . . . . . . . . . . 343 34. . . . . . . . . . . .5 Stochastic calculus and financial markets. . . . . . . . . . . . . . . . . . .13Foreign exchange models.6 Markov processes. . . . . . . . . . . . . 346 35 Notes and References 349 35. . . . . .11Calibration of BGM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353 e 35. . . . . . . . . . 353 35. . . . . . .9 Pricing an interest rate caplet . . . . . . . . . . . . . . . . 350 35. . . . . . . . 345 34. . 350 35. . . . . . . .12Change of num´ raire. . 345 34. . . . 353 35. . . . . . . . . . . .3 Brownian motion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

10 .

the model is used in practice because with a sufficient number of steps. Let us begin with an initial positive stock price S0. Of course. and this will be the case in many of our examples.1 The Binomial Asset Pricing Model The binomial asset pricing model provides a powerful tool to understand arbitrage pricing theory and probability theory. However. within the binomial model we can develop the theory of conditional expectations and martingales which lies at the heart of continuous-time models. we develop notation for the binomial model which is a bit different from that normally found in practice. We denote the price at time 1 by S1 (H ) = uS0 if the toss results in head (H). Secondly.” the price moves down. and by S 1 (T ) = dS0 if it 11 .1) and (1. In the binomial asset pricing model.Chapter 1 Introduction to Probability Theory 1. Thirdly.” the stock price moves up. It is 1 common to also have d = u . d and u. stock price movements are much more complicated than indicated by the binomial asset pricing model. First of all. for what we are about to do we need to assume only (1. In this course. Typically. Let us imagine that we are tossing a coin. assuming that at each step. and change of the stock price from S0 to uS0 represents an upward movement. With this third motivation in mind. it provides a good. the stock price will be either dS 0 or uS0. we take d and u to satisfy 0 < d < 1 < u. we shall use it for both these purposes. and when we get a “Head. but when we get a “Tail. computationally tractable approximation to continuous-time models.1) such that at the next period. We consider this simple model for three reasons. There are two positive numbers. with 0<d<u (1. strictly speaking. we model stock prices in discrete time. the stock price will change to one of two possible values. within this model the concept of arbitrage pricing and its relation to risk-neutral pricing is clearly illuminated. so change of the stock price from S0 to dS0 represents a downward movement.2) below.

1: Binomial tree of stock prices with S 0 results in tail (T). S2 depends on only the first two components of ! . To emphasize this. let us assume that the third toss is the last one and denote by = fHHH HHT HTH HTT THH THT TTH TTT g the set of all possible outcomes of the three tosses. In this case. 1. we have !1 = H . 1 Example 1. we introduce a money market with interest rate r. The stock price Sk at time k depends on the coin tosses. For the moment. and the elements ! of are called sample points. We denote the k-th component of ! by !k . Actually. S2 (HH ) = uS1(H ) = u2 S0 S2(HT ) = dS1(H ) = duS0 S2 (TH ) = uS1(T ) = udS0 S2(TT ) = dS1(T ) = d2S0: After three tosses. for while S 3 depends on all of ! . For example. there are eight possible coin sequences. the price will be one of: = 4. each sample point ! is a sequence of length three. After the second toss. $1 invested in the money market becomes $(1 + r) in the next period. The set of all possible outcomes of a random experiment is called the sample space for the experiment. although not all of them result in different stock prices at time 3. u = 1=d = 2.1 Set S0 = 4. !2 = T and !3 = H . this notation does not quite tell the whole story. Sometimes we will use notation such S 2(!1 !2) just to record more explicitly how S 2 depends on ! = (!1 !2 !3 ). Each sample point ! = (! 1 !2 !3) represents a path through the tree. S 1 depends on only the first component of ! . we often write S k (! ). To complete our binomial asset pricing model. we can think of the sample space as either the set of all possible outcomes from three coin tosses or as the set of all possible paths through the tree. u = 2 and d = 2 .12 S2 (HH) = 16 S (H) = 8 1 S2 (HT) = 4 S =4 0 S1 (T) = 2 S2 (TH) = 4 S2 (TT) = 1 Figure 1. when ! = HTH . Thus. and S0 does not depend on ! at all. We take r to be the interest . We have then the binomial “tree” of possible stock prices shown in Fig.1.

0S0 ): . If the stock goes up. since even in the worst case. In either case. For example. If V0 is more than necessary to buy the 0 shares of stock. You can use the proceeds V0 of the sale of the option for this purpose. in such an application. and so is worth S 1 . and no one would invest in the stock. the stock price rises at least as fast as the debt used to buy it. K )+ if !1 = T . you want to choose V 0 and 0 to also have (1.CHAPTER 1. K )+ if !1 = H and to pay off (dS0 .3) V1(H ) = 0S1 (H ) + (1 + r)(V0 . Our first task is to compute the arbitrage price of this option at time zero. Thus. 0 S0) and you need to have V1(T ). K is positive and is otherwise worth zero. the value of your portfolio (excluding the short position in the option) is With the stock as the underlying asset. you invest the residual money at interest rate r. We denote by V1(!) = (S1(! ) . K at time 1 if S1 . because in a many applications of the model. Of course. and then borrow if necessary at interest rate r to complete the purchase. if 1 + r u. The inequality d 1 + r cannot happen unless either r is negative (which never happens. K )+ = maxfS1(!) . K > 0 and expiration time 1. 0 S0): If the stock goes down. the value of your portfolio is 0 S1 (T ) + (1 + r)(V0 . where 0 is still to be determined. and we can and do sometimes write V1 (!1) rather than V1(! ). At the time you sell the option. you will have V 0 . 0 S0 ) and you need to have V1(H ). let us consider a European call option with strike price 0 S1(H ) + (1 + r)(V0 .4) V 1 (T ) = 0S1 (T ) + (1 + r)(V0 . except maybe once upon a time in Switzerland) or d 1. V 1(! ) actually depends only on ! 1 . You will also own 0 shares of stock. where V0 is still to be determined. it just goes up less than if we had gotten a head. you want to choose V 0 and 0 so that (1.2) The model would not make sense if we did not have this condition.) We assume that d < 1 + r < u: (1. the stock does not really go “down” if we get a tail. This option confers the right to buy the stock at time 1 for K dollars. then the rate of return on the money market is always at least as great as and sometimes greater than the return on the stock. Thus. she should take r to be the rate of interest for her activity. Suppose at time zero you sell the call for V0 dollars. You hedge your short position in the option by buying 0 shares of stock. an agent is either borrowing or lending (not both) and knows in advance which she will be doing. where this quantity might be negative. In the latter case. You now have an obligation to pay off (uS 0 . One should borrow money at interest rate r and invest in the stock. (This is not as ridiculous as it first seems. Introduction to Probability Theory 13 rate for both borrowing and lending. 0S0 dollars invested in the money market. you don’t yet know which value ! 1 will take. K 0g the value (payoff) of this option at expiration.

d (1. this leads to the formula 1 r V0 = 1 + r 1 + . They appeared when we solved the two equations (1. the payoff of this option is V 2 = (S2 . that my definition of 0 is the number of shares of stock one holds at time zero. and (1.9) Because we have taken d < u. both p and q are defined.5) 0= V1(H ) . We want to determine the arbitrage price for this option at time zero. d ~ u d so that (1. they are nothing more than a convenient tool for writing (1. S1 (T )) (1. Depending on how uncertainty enters the model.d This is the arbitrage price for the European call option with payoff V 1 at time 1. (1 + r) V1 (T ) : u d u. This formula is so pervasive the when a practitioner says “delta”. however.9). After some simplification. not the definition of 0 itself.8) 1 ~ V0 = 1 + r pV1(H ) + qV1(T )]: ~ (1. the first and second coin tosses.7) becomes q = u .4). we can regard ~ ~ them as probabilities of H and T . and because they sum to 1. d V1(H ) + u .3) and (1.i.3) and (1. . V1(T ) = so that 0 (S1(H ) . To complete the solution of (1. At expiration. respectively.6) is a consequence of this definition. according to which the number of shares of an underlying asset a hedge should hold is the derivative (in the sense of calculus) of the value of the derivative security with respect to the price of the underlying asset.e. at this point. She then buys 0 shares . formula. we substitute (1. where V0 is still to be determined.7) as (1. where V2 and S2 depend on !1 and !2 . ~ ~ Because of (1.4) from (1.7) To simplify this r p = 1 + .2).4) and solve for V0 .14 These are two equations in two unknowns. Suppose an agent sells the option at time zero for V0 dollars.3) or (1. K )+ . S1(T ) : (1. We now consider a European call which pays off K dollars at time 2. there can be cases in which the number of shares of stock a hedge should hold is not the (calculus) derivative of the derivative security with respect to the price of the underlying asset. V1(T ) S1(H ) . (1 + r) = 1 . Note. and have nothing to do with the actual probabilities of getting H or T on the coin tosses. and we solve them below Subtracting (1. both p and q are in the interval (0 1).8) is not zero. we define (1. she means the derivative (in the sense of calculus) just described.4).. p ~ ~ u.6) into either (1. In fact.3).6) This is a discrete-time version of the famous “delta-hedging” formula for derivative securities. the denominator in (1. They are the risk-neutral probabilites. we obtain V1 (H ) ..

investing V 0 . 0S0 dollars in the money market to finance this. In the next period. we obtain the “delta-hedging for(1. 1 (T ).10) Although we do not indicate it in the notation.10): X1(H ) = X1(T ) = 0S1 (H ) + (1 + r)(V0 . in the six unknowns V 0 .CHAPTER 1. At time 1. and thereby determine the arbitrage price V 0 at time zero of the option and the hedging portfolio 0 .11) Although we do not indicate it in the notation. 1(H ) and 1 (T ). her wealth will be given by the right-hand side of the following equation. the agent has a portfolio (excluding the short position in the option) valued at X1 = 0 S1 + (1 + r)(V0 . 0 .12) Subtracting one of these from the other and solving for mula” 1(T ) = V2 (TH ) . 1 (H )S1(H )) 1(H )S2(HT ) + (1 + r)(X1(H ) . 0 S0 ): After the first coin toss. V2(TH ) = V2(TT ) = 1 (T )S2(TH ) + (1 + r)(X1(T ) . 0 S0): (1. 1(T )S1(T )) 1 (T )S2(TT ) + (1 + r)(X1(T ) . 0 S0) 0S1 (T ) + (1 + r)(V0 . X1 . 1 (T )S1(T )): To solve these equations. 1S1 in the money market. Introduction to Probability Theory 15 of stock. we begin with the last two We now have six equations.10) and the four represented by (1. the agent has X 1 dollars and can readjust her hedge. she wants to have V2 = 1 S2 + (1 + r)(X1 . the outcome of the first coin toss. X1 (H ). Considering all four possible outcomes. 1 (T )S1(T )): 1(T ). S2(TT ) and substituting this into either equation.13) . 1(T )S1(T )) 1(T )S2(TT ) + (1 + r)(X1(T ) . S 1 and therefore X1 depend on !1 . 1(H ). V2(TT ) S2(TH ) . we can solve for 1 ~ X1 (T ) = 1 + r pV2(TH ) + qV2 (TT )]: ~ (1. Therefore. and X1 (T ). Suppose she decides to now hold 1 shares of stock. and she wants it to be V 2. we can write (1. where 1 is allowed to depend on ! 1 because the agent knows what value !1 has taken. She invests the remainder of her wealth.11). the outcomes of the first two coin tosses. the two represented by (1. 1 S1): (1.11) as four equations: V2(HH ) V2 (HT ) V2 (TH ) V2(TT ) = = = = 1(H )S2(HH ) + (1 + r)(X1(H ) . 1(H )S1(H )) 1(T )S2(TH ) + (1 + r)(X1(T ) . S 2 and V2 depend on !1 and !2 . there are really two equations implicit in (1. Thus.

1 of the derivative security. If Vk denotes the value at time k of a derivative security. fHHH HHT HTH HTT g. 1 coin tosses result in the outcomes !1 : : : !k 1 . and itself. 1 tosses !1 : : : !k. .6) and (1. . the portfolio to hedge a short position should hold k.17) H ) .15) and X1(H ) = V1(H ).10). where H ) .14). The first two equations implicit in (1. fTTT g. V : : : !k 1 ) = Sk (!1 :: :: :: !k !k k (!1 . and results again in (1.14) The hedger should choose her portfolio so that her wealth X 1 (T ) if !1 = T agrees with V1(T ) defined by (1.16) 1 ~ V1(H ) = 1 + r pV2(HH ) + qV2(HT )]: ~ This is again analgous to formula (1. 1 ~ Vk 1 (!1 : : : !k 1) = 1 + r pVk (!1 : : : !k . Vk (!1 : : : !k 1 T ) (1. but postponed by one step.11) lead in a similar way to the formulas 1 (H ) = V2(HH ) . . when the first k . and we denote it by V1(T ). 1 1 . This formula is analgous to formula (1.3) and (1. 1 T )] (1. gives the value the hedging portfolio should have at time 1 if the stock goes down between times 0 and 1. . Sk (!1 : : : !k 1 T ) and the value at time k . 1 H ) + q Vk (!1 : : : !k ~ .1) shares of stock.9).4). An example to keep in mind is = fHHH HHT HTH HTT THH THT TTH TTT g (2. The solution of these equations for 0 and V0 is the same as the solution of (1. .1) of all possible outcomes of three coin tosses. The pattern emerging here persists. .1 are known. Some sets in F are . How many sets are there in F ? . We have just shown that 1 ~ V1(T ) = 1 + r pV2 (TH ) + qV2(TT )]: ~ (1. postponed by one step. and this depends on the first k coin tosses ! 1 : : : !k . Finally. 1. where V1(H ) is the value of the option at time 1 if ! 1 = H .9). is given by k 1 (!1 . regardless of the number of periods.13).16 Equation (1.18) 1. V2(HT ) S2(HH ) . defined by (1.1 (!1 : : : !k. we plug the values X 1(H ) = V1(H ) and X1(T ) = V1(T ) into the two equations implicit in (1. after the first k . We define this quantity to be the arbitrage value of the option at time 1 if !1 = T . then at time k .9).2 Finite Probability Spaces Let be a set with finitely many elements. S2(HT ) (1. Let F be the set of all subsets of .

2 Suppose a coin has probability 1 for H and 2 for T .1(ii) in the form (2. A -algebra is a collection following three properties: (i) G of subsets of with the 2 G. Example 1. For the individual elements of 3 3 in (2.1 to determine IP (A) for the remaining sets A 2 F .1 A probability measure properties: (i) 17 IP is a function mapping F into 0 1] with the following IP ( ) = 1. Imagine that is the set of all possible outcomes of some random experiment. We think of IP (A) as this probability. There is a certain probability. (ii) If A1 A2 : : : is a sequence of disjoint sets in F .2 Let be a nonempty set. . Let A be a subset of F . Definition 1. Introduction to Probability Theory Definition 1. that when that experiment is performed. X ! A 2 IP f!g: 2 (2.2) 3 IP fHHH HHT HTH HTT g = 1 + 2 1 3 3 2 + 1 3 3 2 3 2 =1 3 which is another way of saying that the probability of H on the first toss is 1 .1). we define 1 3 1 3 1 3 1 3 3 2 2 3 2 1 3 2 2 3 IP fHHT g = IP fHTT g = IP fTHT g = IP fTTT g = 1 2 2 3 3 1 2 2 3 3 1 2 2 3 3 2 3: 3 IP (A) = For example.2) (see Problem 1. then IP 1 k=1 Ak = ! X 1 k=1 IP (Ak ): Probability measures have the following interpretation. define IP fHHH g = IP fHTH g = IP fTHH g = IP fTTH g = For A 2 F . In the above example. and then used Definition 1. the outcome will lie in the set A.4(ii)) to determine IP for all the other sets in F . 3 As in the above example. between 0 and 1.CHAPTER 1. we specified the probability measure only for the sets containing a single element. it is generally the case that we specify a probability measure on only some of the subsets of and then use property (ii) of Definition 1.

for every set in F 1 whether or not the outcome is in that set. you would be told that the outcome is not in and is in . then its complement Ac (iii) If A1 A2 A3 : : : is a sequence of sets in G . you have been told that the first toss was a T .2: ( ) fHHH HHT HTH HTT g fTHH THT TTH TTT g fHHH HHT g fHTH HTT g fTHH THT g fTTH TTT g ) and all sets which can be built by taking unions of these F3 = F = The set of all subsets of : To simplify notation a bit. 1 Here are some important -algebras of subsets of the set F0 = F1 = F2 = ( ( ) in Example 1. For example. The -algebra F1 is said to contain the “information of the first toss”. F 2 contains the “information of . let us define AH = fHHH HHT HTH HTT g = fH on the first tossg AT = fTHH THT TTH TTT g = fT on the first tossg so that and let us define F1 = f AH AT g AHH = fHHH HHT g = fHH on the first two tossesg AHT = fHTH HTT g = fHT on the first two tossesg ATH = fTHH THT g = fTH on the first two tossesg ATT = fTTH TTT g = fTT on the first two tossesg so that F2 = f AHH AHT ATH ATT AH AT AHH ATH AHH ATT AHT ATH AHT ATT Ac Ac Ac Ac g: HH HT TH TT We interpret -algebras as a record of information. you might be told that the outcome is not in A H but is in A T . Moreover. Similarly. then 2 G. and nothing more. but you are told. In effect. and you are not told the outcome.18 (ii) If A 2 G . Suppose the coin is tossed three times. k=1 Ak is also in G . which is usually called the “information up to time 1”.

1. We have S2(HHH ) = S2(HHT ) = 16 S2(HTH ) = S2(HTT ) = S2(THH ) = S2 (THT ) = 4 S2(TTH ) = S2(TTT ) = 1: Let us consider the interval 4 27]. Then you know that in the first two tosses. Introduction to Probability Theory 19 the first two tosses.3 Let be given by (2. u = 2 and d = 1 . Note that F2 defined earlier contains all the sets which are in (S 2). The “random variable” S0 is really not random.3 Let be a nonempty finite set. called the -algebra generated by the random variable S 2. where S0 = 4.” which is the “information up to time 2. A filtration is a sequence of -algebras F 0 F1 F2 : : : such that each -algebra in the sequence contains all the sets contained by the previous -algebra. and even more.1. and is denoted by (S2). The information content of this -algebra is exactly the information learned by observing S 2 . but someone is willing to tell you. For example. More specifically. This information is the same you would have gotten by being told that the value of S 2(! ) is 4. Knowing whether the outcome ! of the three tosses is in (it is not) and whether it is in (it is) tells you nothing about ! Definition 1. In particular. Fn F be the -algebra of all subsets of . Nonetheless. but you cannot make this distinction from knowing the value of S2 alone. and is not in A TT . This collection of sets is a -algebra. for each set in (S 2). A random variable maps into IR.1) and consider the binomial asset pricing Example 1. 2 S2 (HHT ) = u2 S0 = 16. Definition 1. if you see the first two tosses. whether ! is in the set. you can distinguish A HT from ATH . You might be told.4 Let be a nonempty finite set and let random variable is a function mapping into IR. is in AHT ATH . the random variable S2 of Example 1. This means that the information in the first two tosses is greater than the information in S 2 . there was a head and a tail. The so-called “trivial” -algebra F 0 contains no information. S2 and S3 are all random variables. and we can look at the preimage under the random variable of sets in IR.CHAPTER 1. for example. S1. albeit a degenerate one. since S0(!) = 4 for all ! 2 . suppose the coin is tossed three times and you do not know the outcome ! . that ! is not in AHH . A Example 1.” The -algebra F 3 = F contains “full information” about the outcome of all three tosses. for example. . and you know nothing more. The preimage under S2 of this interval is defined to be we can get as preimages of sets in IR is: f! 2 S2(!) 2 4 27]g = f! 2 4 S2 27g = Ac : TT The complete list of subsets of AHH AHT ATH ATT and sets which can be built by taking unions of these. Then S0. it is a function mapping into IR. and thus technically a random variable. Consider.

S 2 (TTH ) = S2(TTT ) = 1. Given any set A IR. we define the induced measure of A to be LX (A) = IP fX 2 Ag: In other words. in which case the induced measure of a set A IR is the integral of the density over the set A. as in the case of S2 above. nothing more.20 Definition 1. in the discussion above.5 Let be a nonemtpy finite set and let F be the -algebra of all subsets of . the induced measure of S2 places a mass of size 1 3 FS2 (x) = IP (S2 By the distribution of a random variable 8 >0 >4 <9 x) = > 8 >1 :9 if x < 1 if 1 x < 4 if 4 x < 16 if 16 x: (2. We say that X is G -measurable if every set in (X ) is also in G . In order to work through the concept of a risk-neutral measure. (Later we will consider random variables X which have densities. we can either tell where the masses are and how X . A common way to record this 9 3 9 information is to give the cumulative distribution function F S2 (x) of S2 . where A is a subset of IR. let IP be a probabilty measure on ( F ). It has an existence quite apart from discussion of probabilities.-algebra of F . defined by 2 In fact. Definition 1. finite set. let F be the -algebra of all subsets of .) LX . The -algebra (X ) generated by X is defined to be the collection of all sets of the form f! 2 X (! ) 2 Ag. If X is discrete.6 Let be a nonempty. Let X be a random variable on ( F ). Note: We normally write simply fX 2 Ag rather than f! 2 X (!) 2 Ag. regardless of whether the probability for H is 1 or 1 . we mean any of the several ways of characterizing Important Note. 3 2 . we set up the definitions to make a clear distinction between random variables and their distributions. or tell what the cumulative distribution function is. the induced measure of a set A tells us the probability that X takes a value in A.2. some sets in IR and their induced measures are: LS2 ( ) = IP ( ) = 0 LS2 (IR) = IP ( ) = 1 LS2 0 1) = IP ( ) = 1 2 LS2 0 3] = IP fS2 = 1g = IP (ATT ) = 2 : 3 1 = 9 at the number 16. For example. In the case of S2 above with the probability measure of Example 1. Let G be a sub. a mass of size 4 at the number 4. A random variable is a mapping from to IR. and let X be a random variable on .3) large they are. and a mass of size 2 2 = 4 at the number 1.

CHAPTER 1. the payoff of the put is 2 with probability 1 and 0 with probability 3 . If we set the probability 3 9 of H to be 1 .1. and let X be a random variable on . and the probability it is zero is 3 .4) as IEX = = = = = X 2 ! n X X (!)IP f!g X 2f g k=1 ! Xk =xk n X X xk IP f!g k=1 ! Xk =xk n X xk IP fXk = xk g k=1 n X xk LX fxk g: k=1 2f g X (!)IP f! g . let F be the -algebra of all subsets of . Since finite set. i. It depends on the random variable X and the probability measure IP we use in . The distribution of S 2 has changed. If we set the probability of H to be 1 . Consider also 4 4 a European put with strike price 3 expiring at time 2. finite set. let IP be a probabilty measure on ( F ). In a similar vein. Suppose in the binomial model of Example 1. Like the payoff of the call. two different random variables can have the same distribution. a way of assigning probabilities to sets in IR.. X can take only finitely many values. The expected value of X is defined to be IEX = X 2 ! X (!)IP f!g: (2. which takes the value 2 if ! = TTH or ! = TTT . then LS2 assigns mass 1 to the number 16. 14)+.e. the probability of H and the probability of T is 1 .7 Let be a nonempty. The probability the payoff is 2 is 1 . which we label x 1 : : : xn . then LS2 assigns mass 1 to the number 16. The payoff of the put at time 2 is (3 . and then rewrite (2. and a “risk-neutral” distribution). The payoff of the call at time 2 is the random variable (S2 . and takes the value 0 in every other case. We can partition the subsets fX 1 = x1 g : : : fXn = xn g. Introduction to Probability Theory 21 The distribution of a random variable is a measure L X on IR. S 2)+ . Consider a 2 European call with strike price 14 expiring at time 2. The payoffs of the call and the put are different 4 4 random variables having the same distribution. It is still defined by S2(HHH ) = S2(HHT ) = 16 S2(HTH ) = S2(HTT ) = S2(THH ) = S2 (THT ) = 4 S2(TTH ) = S2(TTT ) = 1: Thus. Definition 1. but 2 4 the random variable has not.4) is a into Notice that the expected value in (2. which takes the value 2 if ! = HHH or ! = HHT . a random variable can have more than one distribution (a “market” or “objective” distribution.4) is defined to be a sum over the sample space .

The variance of X is defined to be the expected value of (X . finite set. IEX )2IP f! g: (2.5) x1 : : : xn. IEX )2. We need this integral because. The definition of IES2 is IES2 = S2 (HHH )IP fHHH g + S2(HHT )IP fHHT g +S2 (HTH )IP fHTH g + S2(HTT )IP fHTT g +S2 (THH )IP fTHH g + S2(THT )IP fTHT g +S2 (TTH )IP fTTH g + S2 (TTT )IP fTTT g = 16 IP (AHH ) + 4 IP (AHT ATH ) + 1 IP (ATT ) = 16 IP fS2 = 16g + 4 IP fS2 = 4g + 1 IP fS2 = 1g = 16 LS2 f16g + 4 LS2 f4g + 1 LS2 f1g = 16 1 + 4 4 + 4 4 9 9 9 48 : = 9 Definition 1. a generalization of the Riemann integral. although the expected value is defined as a sum over the sample space . The collection of subsets of IR we consider. and for which Lebesgue measure is defined. i. IEX )2IP fX = xk g = n X k=1 (xk . if X takes the values Var(X ) = n X k=1 (xk . We define the Lebesgue measure of intervals in IR to be their length.e. the generalization to other spaces will be given later. we consider the set of real numbers IR. subsets of IR. Indeed. let IP be a probabilty measure on ( F ). To make the above set of equations absolutely clear.3). then One again.22 Thus. IEX )2LX (xk ): 1. it can be defined on abstract spaces. unlike the Riemann integral. we can also write it as a sum over IR. such as the space of infinite sequences of coin tosses or the space of paths of Brownian motion. We use Lebesgue measure to construct the Lebesgue integral. we can rewrite (2. which is uncountably infinite. . but not all.3 Lebesgue Measure and the Lebesgue Integral In this section.5) as a sum over IR rather than over . This definition and the properties of measure determine the Lebesgue measure of many. Var(X ) = X 2 ! (X (! ) .8 Let be a nonempty. is the collection of Borel sets defined below. and let X be a random variable on . we consider S 2 with the distribution given by (2. let F be the -algebra of all subsets of . This section concerns the Lebesgue integral on the space IR only..

Since B(IR) is a -algebra. every complement of a Borel set is Borel. since they can be written as intersections of open half-lines and closed half-lines. so B(IR) contains a b] = (. is the smallest -algebra containing all open intervals in IR. as is 1 n=1 1 (a a + n) (a . every union of open intervals is also in B(IR). for every real number a. By definition. For example.1 b] \ (a 1): Every set which contains only one real number is Borel. The following discussion of this fact uses the -algebra properties developed in Problem 1. Introduction to Probability Theory 23 Definition 1. n a+ n : = fa 1 a2 : : : an g. The sets in B(IR) are called Borel sets.3. Every set which can be written down and just about every set imaginable is in B(IR). if a is a real number.1 a) (b 1) is Borel. the closed half-lines c a 1) = and 1 n=1 1 a a + n] a . n a): (. (a b] = (. if A then A= k=1 fak g: . every open interval (a b) is in B(IR). In addition. Indeed. where a and b are real numbers. n a] (. denoted B(IR). the union n=1 (. n This means that every set containing finitely many real numbers is Borel. Since B(IR) is a -algebra.1 a) = For real numbers a and b.CHAPTER 1.9 The Borel -algebra. then fag = \ 1 n=1 1 1 a.1 a] = n=1 are Borel.1 a) (b 1) : This shows that every closed interval is Borel. the open half-line (a 1) = is a Borel set. Half-open and half-closed intervals are also Borel. For example.

remove the open set The remaining set length 41 . We have just seen that any non-Borel set must have uncountably many points. the length of the 8 8 8 .24 In fact. remove the middle half.) This example gives a hint of how complicated a Borel set can be. From each of these pieces. so at stage k.. 3 1 3 13 1 0 1 4 1 16 16 16 15 64 : : : 4 16 . the set C k has 2k pieces. the total length removed is k=1 X1 2k = 1 1 and so the Cantor set. We shall see eventually that the Cantor set has uncountably many points.” Despite the fact that the Cantor set has no “length. The “length” of A2 . Continue this process.e. Note that the length of A 1 . as is its complement. i.. Thus.e. Example 1. the first set removed. then k=1 fak g: This means that the set of rational numbers is Borel. 2 1 1 is 1 + 1 = 4 . is 1 . remove the open interval 3 A1 = 1 4 : 4 The remaining set C1 = 0 1 4 1 3 A2 = 16 16 31 4 13 15 : 16 16 has two pieces. In particular. the points are all in C . There are. however. The Cantor set k 15 16 1 : has four pieces. every set containing countably infinitely many numbers is Borel.4 (The Cantor set. the set of irrational numbers. the second set removed. none of the endpoints of the pieces of the sets C 1 C2 : : : is ever removed. if A = fa 1 A= n a2 : : : g. the set of points not removed. and each piece has 1 C2 = 0 16 3 1 16 4 3 13 4 16 C= \ 1 k=1 Ck is defined to be the set of points not removed at any stage of this nonterminating process. Consider the unit interval 0 1]. and remove the middle half.k k-th set removed is 2 . sets which are not Borel. i. Thus. has zero “length. This is a countably infinite set of points. and in general. The “length” of the next set removed is 4 32 = 1 .” there are lots of points in this set. We use it later when we discuss the sample space for an infinite sequence of coin tosses.

n a + n = n: 0 fag = 0: . and property (v) in Problem 1. except that the total measure of the space is not necessarily 1 (in fact. so 0 (\1 Ak ) = 0. (ii) If A1 A2 : : : is a sequence of disjoint sets in B(IR). then (A 1 ) < 1 is needed in (v). then 1 k=1 Ak = ! X 1 k=1 (Ak ): Lebesgue measure is defined to be the measure on (IR B(IR)) which assigns the measure of each interval to be its length. we denote Lebesgue measure by 0 . the Lebesgue measure of the Cantor set in Example 1. we must have 1 1 a. For example. n a+ n 0 0 Letting n ! 1.4 must be zero. and that determines the Lebesgue measure of all other Borel sets. Following Williams’s book. (A) in Problem 1. A measure on (IR function mapping B into 0 1] with the following properties: (i) 25 B(IR)) is a ( ) = 0. k=1 We specify that the Lebesgue measure of each interval is its length.4 needs to be modified to say: (v) If A1 A2 : : : is a sequence of sets in B(IR) with A 1 A2 \ 1 k=1 To see that the additional requirment Ak = nlim (An ): !1 ! and (A1 ) < 1.CHAPTER 1. A measure has all the properties of a probability measure given in Problem 1. one no longer has the equation (Ac ) = 1 . The Lebesgue measure of a set containing only one point must be zero.4(iii). 0 (IR) = 1). since fag for every positive integer n.4. In fact. because of the “length” computation given at the end of that example.10 Let B(IR) be the -algebra of Borel subsets of IR. but limn!1 0 (An ) = 1. Introduction to Probability Theory Definition 1. consider A1 = 1 1) A2 = 2 1) A3 = 3 1) : : :: Then \1 Ak k=1 = . we obtain 0 fag 1 1 2 a .

a function of the form h(x) = where each gk is of the form k=1 ck gk (x) if x 2 Ak if x 2 Ak = gk (x) = hd 0 = ck ( and each ck is a real number. and we shall pretend such functions don’t exist. Z IR f d 0 = sup Z IR hd 0 h is simple and h(x) f (x) for every x 2 IR : . Definition 1. The integral was invented to get around this problem. We may not compute the Lebesgue measure of an uncountable set by adding up the Lebesgue measure of its individual members. i.12 An indicator function g from and 1. Definition 3. We call f be a function from IR to IR. we want the -algebra generated by f to be contained in B(IR). because there is no way to add up uncountably many numbers. It is difficult to conceive of a function which is not Borel-measurable.e. Indeed. Hencefore. then 0 (A) = X 1 k=1 0 fak g = X 1 k=1 0 = 0: The Lebesgue measure of a set containing uncountably many points can be either zero. or infinite.26 The Lebesgue measure of a set containing countably many points must also be zero.. “function mapping IR to IR” will mean “Borel-measurable function mapping IR to IR” and “subset of IR” will mean “Borel subset of IR”. Definition 1. In order to think about Lebesgue integrals. possibly taking the value 1 at some points.11 Let fx 2 IR f (x) 2 Ag is in B(IR) whenever A 2 B(IR). We define the Lebesgue integral of h to be 1 0 Z n X Z R k=1 IR gk d 0 = n X k=1 ck 0 (Ak ): We Let f be a nonnegative function defined on define the Lebesgue integral of f to be IR. we must first consider the functions to be integrated. positive and finite. if A = fa1 a2 : : : g.4 is purely technical and has nothing to do with keeping track of information. We say that f is Borel-measurable if the set IR to IR is a function which takes only the values 0 A = fx 2 IR g (x) = 1g the set indicated by g . In the language of Section 2. We define the Lebesgue integral of g to be Z IR g d 0 = 0 (A): n X A simple function h from IR to IR is a linear combination of indicators.

1 xn].. 1. and we also approximate it from below by the lower sum f (rk )(xk . Let A be a subset of IR. Introduction to Probability Theory It is possible that this integral is infinite.1 at other points. 0 (xk . we say that f is integrable. IR jf j d 0 < 1. ( IR I l Af d 0 1 0 if x 2 A if x 2 A = The Lebesgue integralRjust defined is related to the Riemann integral in one very important way: if R b the Riemann integral a f (x)dx is defined. xk 1 ) = . possibly taking the value 1 at some points and the value . We approximate the Riemann integral from above by the upper sum R 0 1] f d 0 = 0: n X k=1 n X k=1 f (qk )(xk . we choose partition points 0 = x 0 < x1 < < xn = 1 and divide the interval 0 1] into subintervals x 0 x1] x1 x2] : : : xn. R 0 R Z A where fd 0= Z I l A (x) = is the indicator function of A. d 0 are finite R (or equivalently. let f be a function defined on IR. xk 1 ) = . Example 1. possibly taking the value 1 at some points and the value . Z Let f be a function defined on IR. since jf j = f + + f . where f (qk ) = 1.5 Let Q be the set of rational numbers in 0 1].CHAPTER 1. The Lebesgue integral has two important advantages over the Riemann integral. In each subinterval xk. and we define the Lebesgue integral of f to be . Q has Lebesgue measure zero. Z IR f d . xk 1 ) = 1 . then the Lebesgue integral a b] f d 0 agrees with the Riemann integral. If it is finite. xk 1 ) = 0: . Being a countable I 0 1] is 1 To compute the Riemann integral 0 f (x)dx. and there is also an irrational point rk .1 at other points. we say that f is integrable. ).f (x) 0g respectively. If both IR f + d 0 and IR f . and so the Lebesgue integral of f over Z = l Q .1 xk ] there is a rational point q k . where f (rk ) = 0. We define provided the right-hand side is not of the form 1 . n X k=1 n X k=1 1 (xk . 27 Finally. We define the positive and negative parts of f to be f + (x) = maxff (x) 0g f (x) = maxf. and consider f set. . IR fd 0= Z IR f+ d 0. The first is that the Lebesgue integral is defined for more functions. as we show in the following examples.

Example 1. Thus the upper approximating sum is 1. the upper sum is always 1 and the lower sum is always 0.1 f (x) dx. and thus has Lebesgue integral Z IR h d 0 = y 0 f0 g = 0 : It follows that Z IR f d 0 = sup Z IR hd 0 h is simple and h(x) f (x) for every x 2 IR = 0: 1 Now consider the Riemann integral . this point will contribute 1 times the length of the subinterval containing it. one of these will contain Riemann integral . and when we compute the upper approximating sum for .1 f (x) dx. Every simple h(x) = for some y ( 0 y if x = 0 if x 6= 0 2 0 1). Since these two do not converge to a common value as the partition becomes finer. we have IR cf d = c IR Z fd 0+ fd 0 Z IR gd 0 IR Z Finally. which for this function f is the same as the R 1 f (x) dx. then IR fd 0 Z IR gd d 0: Z A B fd 0= Z A fd 0+ Z B f d 0: .6 Consider the function f (x) = ( 0 1 if x = 0 if x 6= 0: This is not a simple function because simple function cannot take the value function which lies between 0 and f is of the form 1. the Riemann integral is not defined.28 No matter how fine we take the partition of 0 1]. the lower approximating sum is 0. for any two functions f and g and any real constant c. R The Lebesgue integral has all linearity and comparison properties one would expect of an integral. Z IR and whenever f (x) (f + g ) d Z 0 0 = Z g (x) for all x 2 IR. and again the Riemann integral does not exist. if A and B are disjoint sets.1 1] into subintervals. In particular. When we partition .1 R1 the point 0. On the other hand.

1 (Fatou’s Lemma) Let fn verging pointwise to a function f .8 Consider a sequence of normal densities. but IR f d 0 = 0. where nlim IR fn d 0 = 1 !1 Z . Example 1.7 Consider a sequence of normal densities. Then R 0 1 if x = 0 if x 6= 0: R R n = 1 2 : : : be a sequence of nonnegative functions con0 Z If limn!1 IR fn d 0 is defined. Theorem 3.CHAPTER 1. but R f d = 0. Before we state the theorems. Introduction to Probability Theory 29 There are three convergence theorems satisfied by the Lebesgue integral. xn : 2 These converge pointwise to the function f (x) = ( We have again IR fn d 0 = 1 for every n. each with mean 0 and the n-th having vari1 ance n : r 2 2 fn (x) = n e.n)2 1 and the n-th having fn (x) = p e 2 2 : These converge pointwise to the function R We have IR fn d 0 = 1 for every n. (x.6 to be zero. the Lebesgue integral of this function was already seen in Example 1. In each of these the situation is that there is a sequence of functions f n n = 1 2 : : : converging pointwise to a limiting function f . Pointwise convergence just means that nlim fn (x) = f (x) for every x 2 IR: !1 There are no such theorems for the Riemann integral.8.7 and 1. so limn!1 IR fn d 0 = 1. because the Riemann integral of the limiting function f is too often not defined. The function f is not the Dirac delta. then Fatou’s Lemma has the simpler conclusion R IR fd fd lim inf n !1 Z IR fn d 0 : Z IR 0 nlim Z !1 IR fn d 0: This is the case in Examples 1. so lim n f (x) = 0 for every x 2 IR: R f d = 1. each with variance mean n: 1 . 0 IR IR n 0 !1 Example 1. we given two examples of pointwise convergence which arise in probability theory.

the smaller one. i. is n=1 R lim inf n!1 IR gn d 0 and the larger one.. 1. Assume that 0 f1 (x) f2 (x) f3 (x) Then Z IR where both sides are allowed to be 1. R f but gn = 2R n if n is odd. which may take either positive or negative values.e. IR g d 0 < 1) such that Then jfn(x)j g(x) for every x 2 IR for every n: Z Z IR f d 0 = nlim !1 IR fn d 0 and both sides will be finite. Assume R that there is a nonnegative integrable function g (i. a probability measure on ( F ). called the sample space. 2.3 (Dominated Convergence Theorem) Let fn n = 1 2 : : : be a sequence of functions.13 A probability space ( (i) (ii) (iii) F IP ) consists of three objects: .4 General Probability Spaces Definition 1. a nonempty set. a -algebra of subsets of .2 (Monotone Convergence Theorem) Let fn converging pointwise to a function f . converging pointwise to a function f . is lim supn!1 IR gn d 0 .e. a function which assigns to each set A 2 F a number IP (A) 2 0 1]. R By definition. Theorem 3. F. which contains all possible outcomes of some random experiment. The key assumption in Fatou’s Lemma is that all the functions take only nonnegative values. which represents the probability that the outcome of the random experiment lies in the set A.7 or 1.8 Rby setting g n = fn if n is even..30 while IR f d 0 = 0. f d 0 = nlim Z IR !1 fn d 0 Theorem 3. IP . but IR gn d 0 = 2 if n is odd. Now IR gn d 0 = 1 if n is even. 1 and 2. We could modify either Example 1. Fatou’s Lemma does not assume much but it is is not very satisfying because it does not conclude that R Z IR f d 0 = nlim Z !1 IR fn d 0: n = 1 2 : : : be a sequence of functions for every x 2 IR: There are two sets of assumptions which permit this stronger conclusion. . 1. The sequence f IR gn d 0 g1 has two cluster points. Fatou’s Lemma guarantees that even the smaller cluster point will be greater than or equal to the integral of the limiting function.

a number between zero and one. We will use this space quite a bit.1). (b) (Countable additivity) If A1 A2 : : : is a sequence of disjoint sets in F . Then the probability of T is q = 1 . and so give it a name: n . We repeat these and give some examples of infinite probability spaces as well.9 Finite coin toss space.) If A1 k=1 1 Ak = nlim IP (An ): !1 ! . then IP 1 k=1 Ak = ! X 1 k=1 IP (Ak ): (c) (Finite additivity) If n is a positive integer and A 1 IP (A1 (d) If A and B are sets in F and A : : : An are disjoint sets in F . we define IP f! g = pNumber of H in ! q Number of T in ! : IP (A) = X ! A 2 IP f!g: (4.1) We can define IP (A) this way because A has only finitely many elements.4 that a probability measure IP has the following properties: (a) IP ( ) = 0. so that is the set of all sequences of H and T which have n components. For each ! = (!1 !2 : : : !n ) in n . p.1 We recall from Homework Problem 1. .) If A1 A2 : : : is a sequence of sets in F with A 1 A2 IP \ ! . then IP (B ) = IP (A) + IP (B n A): IP (B ) IP (A): A2 : : : is a sequence of sets in F with A 1 A2 IP 1 In particular. Introduction to Probability Theory 31 Remark 1. Toss a coin n times. Example 1. Let F be the collection of all subsets of n . Suppose the probability of H on each toss is p. and so only finitely many terms appear in the sum on the right-hand side of (4. (d) (Continuity from below. we define For each A 2 F . then An ) = IP (A1 ) + + IP (An ): B. then (d) (Continuity from above.CHAPTER 1. then k=1 Ak = nlim IP (An ): !1 We have already seen some examples of finite probability spaces.

However. we define Fn to be the -algebra determined by the first n tosses. where these sets were defined earlier. the set containing the single sequence Because sets. . the probability of a H on the second toss is p. . fH on every tossg = is also in F . we must also put into it the sets . For example. F2 is a -algebra. and we need to exercise some care in the construction of the -algebra we will use here. for each positive integer n. then the description of A depends on only the first n tosses. and all unions of the four basic fHHHHH g = fH on every tossg is not in any of the F n -algebras. so that is the set of all nonterminating sequences of H and T .10 Infinite coin toss space. This is an uncountably infinite space. where n ranges over the positive integers. We call this space 1 . p for T . F2 contains four basic sets. AHH = f! = (!1 !2 !3 : : : ) !1 = H !2 = H g = The set of all sequences which begin with HH AHT = f! = (!1 !2 !3 : : : ) !1 = H !2 = T g = The set of all sequences which begin with HT ATH = f! = (!1 !2 !3 : : : ) !1 = T !2 = H g = The set of all sequences which begin with TH ATT = f! = (!1 !2 !3 : : : ) !1 = T !2 = T g = The set of all sequences which begin with TT: In the -algebra F . and it is clear how to define IP (A). suppose A = AHH ATH . Therefore. Then A is in F2. We also put in every other set which is required to make F be a -algebra. Toss a coin repeatedly without stopping. For example. and then we have IP (A) = IP (AHH ATH ) = p2 + qp = (p + q )p = p: In other words. For example. \ 1 n=1 fH on the first n tossesg We next construct the probability measure IP on ( 1 F ) which corresponds to probability p 2 0 1] for H and probability q = 1 . we put every set in every -algebra Fn . because it depends on all the components of the sequence and not just the first n components. For each positive integer n. Let A 2 F be given.32 Example 1. We set IP (AHH ) = p2 and IP (ATH ) = qp. If there is a positive integer n such that A 2 Fn . the set fH on the first n tossesg is in Fn and hence in F .

X (TTTT ) = 0. 2 1 . We define a “dyadic rational m number” to be a number of the form 2k . For example. the only sets which can have positive probabilty in 1 are those which contain uncountably many elements. Of course. Yk (!) = and we also define the random variable ( Y2 : : : by 1 0 if !k if !k =H =T X (! ) = n X Y k (! ) k=1 2k : Since each Yk is either zero or one. X takes values in the interval 0 1]. X (HHHH ) = 1 and the other values of X lie in between. we define a sequence of random variables Y1 If p = 1. then every set in 1 which contains only one element (nonterminating sequence of H and T ) has probability zero.1. otherwise. and then use the properties of probability measures listed in Remark 1.1) which are not dyadic rationals correspond to a single ! have a unique binary expansion. We are in a case very similar to Lebesgue measure: every point has measure zero. Indeed. where k and m are integers.1(d) (continuity from above). For example. Every dyadic rational in (0.1) corresponds to two sequences ! 2 1 . IP fH on every tossg = nlim IP fH on the first n tossesg = nlim pn : !1 !1 A similar argument shows that if 0 < p < 1 so that 0 < q < 1. To determine the probability of these sets. but sets can have positive measure.CHAPTER 1. IP fH on every tossg = 0. then IP fH on every tossg = 1. 3 is a dyadic 4 rational. For example. we write them in terms of sets which are in Fn for positive integers n. In the infinite coin toss space. and hence very set which contains countably many elements also has probabiliy zero. X (HHTTTTT ) = X (HTHHHHH ) = 3: 4 The numbers in (0. these numbers . Introduction to Probability Theory 33 Let us now consider a set A 2 F for which there is no positive integer n such that A 2 F . Such is the case for the set fH on every tossg. fH on the first tossg fH on the first two tossesg fH on the first three tossesg and \ 1 n=1 fH on the first n tossesg = fH on every tossg: According to Remark 1.

2. if we set p = q = 1 in the construction of this example. i. but HT is not. we are removing exactly those points in 0 1] which were removed in the Cantor set construction of Example 3. HHTTTTHH is the beginning of a sequence in pairs . none of these numbers is in C 0.2. we have a corresponding induced measure LX on 0 1]. but the sequence ! must begin with TTTH or TTHT .2. Consider now the set of real numbers C = fX ( ! ) ! 2 0 pairs g: The numbers between ( 1 1 ) can be written as X (! ). Similarly.. For example. by requiring consecutive coin tosses to be paired. m LX m2. 1 2k = 21k : k We conclude this example with another look at the Cantor set of Example 3. but the sequence ! must begin with either 4 2 1 3 TH or HT . one can convince onself that in fact C 0 = C . With a bit more work. .e. The only way this can be is for LX to be Lebesgue measure. so none of these numbers is in C 0. then we have 2 1 LX 0 2 = IP fFirst toss is T g = 1 2 1 1 = IP fFirst toss is H g = 1 LX 2 2 1 LX 0 4 = IP fFirst two tosses are TT g = 1 4 1 LX 4 1 = IP fFirst two tosses are TH g = 1 2 4 1 LX 2 3 = IP fFirst two tosses are HT g = 1 4 4 3 LX 4 1 = IP fFirst two tosses are HH g = 1 : 4 0 Continuing this process.34 Whenever we place a probability measure IP on ( F ). In other words. Let pairs be the subset of in which every even-numbered toss is the same as the odd-numbered toss immediately preceding it. For example. Continuing this process. Therefore. the LX -measure of all intervals in 0 1] whose endpoints are dyadic rationals is the same as the Lebesgue measure of these intervals. the numbers between ( 16 16 ) can be written as X (! ). we can verify that for any positive integers k and m satisfying m.1 < m 2k 2k 1 we have In other words. It is interesing to consider what L X would look like if we take a value of p other than 1 when we 2 construct the probability measure IP on . we see that C 0 will not contain any of the numbers which were removed in the construction of the Cantor set C in Example 3. C 0 C.

CHAPTER 1. Define the standard normal density '(x) = p1 e 2 : 2 .12 provide probability spaces which are often useful approximations to reality. and if the set has uncountably many points. Example 1. Because 0 1] = 1. just as I know of no way to toss a coin infinitely many times. F to be B(IR). Example 1. perhaps using a random number generator. both Examples 1. Nonetheless. 4 or 16. then this probability can be positive.11 Suppose we choose a number from IR in such a way that we are sure to get either 1. For example. the measure induced by the stock price S2 . this requires that every number have probabilty zero of being chosen. we can speak of the probability that the number chosen lies in a particular set. Introduction to Probability Theory 35 In addition to tossing a coin. Let = 0 1] and let F = B( 0 1]). We describe this random experiment by taking 9 9 to be IR. define IP (A) = Z A ' d 0: (4.2) . the collection of all Borel subsets containined in 0 1]. Furthermore. F = B(IR) and for every Borel set A IR. For each Borel set A 0 1]. Here are some probability spaces which correspond to different ways of picking a number at random. another common random experiment is to pick a number. because this interval contains the numbers 1 and 4. the probability of the interval (0 is 8 . 0 This probability space corresponds to the random experiment of choosing a number from 0 1] so that every number is “equally likely” to be chosen. Since there are infinitely mean numbers in 0 1]. and setting up the probability measure so that 4 1 IP f1g = 9 IP f4g = 4 IP f16g = 9 : 9 This determines IP (A) for every set A 2 B(IR). 9 5] The probability measure described in this example is L S2 . the probability of 9 getting 4 is 4 and the probability of getting 16 is 1 . when the initial stock price S 0 = 4 and the probability of H is 1 . this gives us a probability measure.12 Uniform distribution on 0 1]. I know of no way to design a physical experiment which corresponds to choosing a number at random from 0 1] so that each number is equally likely to be chosen. x2 Let = IR.13 Standard normal distribution. we define IP (A) = 0 (A) to be the Lebesgue measure of the set. Example 1.10 and 1. Nonetheless. but not the number 16. This distribution was discussed 3 immediately following Definition 2. we construct the experiment so that the probability of getting 1 is 4 .8.

a 2 This corresponds to choosing a point at random on the real line. but if a set A is given.e. With this sequence.14 Let ( F IP ) be a probability space.. X (! ) = l A (!) = I for some set A 2 F . we can define X dIP = nlim Z !1 Yn dIP: . and let X be a random variable on this space. i. a mapping from to IR. We repeat this construction below. If X is an indicator.2) as the less mysterious Riemann integral: Z b 1 x2 IP a b] = p e 2 dx: . i. then we can write (4. ( 1 0 if ! if ! 2A 2 Ac Z X dIP = IP (A): n X k=1 X (! ) = n X Z ck l Ak (!) I n X where each ck is a real number and each Ak is a set in F . The construction of the integral in a general probability space follows the same steps as the construction of Lebesgue integral.e. we define If X is a simple function.2) is an interval a b]. we define Z X dIP = k=1 ck I l Ak dIP = k=1 ck IP (Ak ): If X is nonnegative but otherwise general. then the probability the point is in that set is given by (4. Definition 1. we define Z X dIP = sup Z Y dIP Y is simple and Y (! ) X (!) for every ! 2 : In fact.2). we can always construct a sequence of simple functions Yn and Y (! ) = limn!1 Yn (! ) for every ! n = 1 2 : : : such that 0 Y1 (! ) Y2 (! ) Y3 (! ) : : : for every ! 2 Z 2 . possibly also taking the values 1. i.36 If A in (4. and every single point has probability zero of being chosen.e.

X (!) 0g Z X dIP = Z X + dIP .e. we don’t need to have X (! ) Y (! ) for every ! 2 in order to reach this conclusion. In particular.4 (Fatou’s Lemma) Let Xn n = 1 2 : : : be a sequence of almost surely nonnegative random variables converging almost surely to a random variable X . we say it holds almost surely. if X and Y are random variables and c is a real constant. where then we define 37 Z X + dIP < 1 Z . . we define Z The expectation of a random variable X is defined to be A X dIP = IEX = Z I l A X dIP: X dIP: Z The above integral has all the linearity and comparison properties one would expect. almost surely is enough. then Z (X + Y ) dIP = Z Z cX dIP = c X dP . X dIP: . then Z X dIP + Z Y dIP If X (! ) Y (!) for every ! 2 Z X dIP Z Y dIP: In fact. then Z A B X dIP = Z A X dIP + Z B X dIP: We restate the Lebesgue integral convergence theorem in this more general context. X dIP < 1 . it is enough if the set of ! for which X (! ) Y (! ) has probability one. X + (!) = maxfX (!) 0g X (! ) = maxf. i. When a condition holds with probability one. We acknowledge in these statements that conditions don’t need to hold for every ! . Then Z X dIP lim inf n !1 !1 Z Xn dIP or equivalently. if A and B are disjoint subsets of and X is a random variable. Z If A is a set in F and X is a random variable.CHAPTER 1. Finally. Introduction to Probability Theory If X is integrable. Theorem 4. IEX lim inf IEXn: n .

4) The probability measure IP weights different parts of the real line according to the density '.5) . the market measure and the risk-neutral measures in financial markets are related this way. whenever ' is a nonnegative function defined on R satisfying IR ' d 0 = 1. converging almost surely to a random variable X . we constructed a probability measure on (IR B(IR)) by integrating the standard R normal density.3).3) We shall often have a situation in which two measure are related by an equation like (4.38 Theorem 4.13. Assume that 0 X1 X2 X3 Then almost surely: Z X dIP = nlim Z !1 Xn dIP or equivalently. IEX = nlim IEXn : !1 Theorem 4. Assume that there exists a random variable Y such that jXnj Y almost surely for every n: Then Z X dIP = nlim Z !1 Xn dIP or equivalently. Now suppose f is a function on (R B(IR) IP ).5 (Monotone Convergence Theorem) Let Xn n = 1 2 : : : be a sequence of random variables converging almost surely to a random variable X .3) is the Radon-Nikodym derivative of dIP with respect to 0 . Definition 1. In fact. We say that ' in (4. and we write dIP '= d : 0 (4. We want to show that IR f dIP = Z IR f' d 0 (4. IEX = nlim IEXn : !1 In Example 1.6 (Dominated Convergence Theorem) Let Xn n = 1 2 : : : be a sequence of random variables.14 gives us a value for the abstract integral Z We can also evaluate Z Z IR IR f dIP: 0 f' d which is an integral with respec to Lebesgue measure over the real line. we call ' a density and we can define an associated probability measure by IP (A) = Z A 'd 0 for every A 2 B(IR): (4. In fact.

5) holds when f is a simple function.4) (substitute d 0 for ' in (4. Now that we know that (4. f (x) that case. Then Z IR f dIP = = = = = Z "X n ck hk dIP IR k=1 n X Z ck hk dIP IR k=1 Z n X ck hk ' d 0 k=1 " IR # Z X n ck hk ' d 0 IR k=1 Z IR # f' d 0: Step 3. is a f (x) = n X k=1 ck hk (x) where each ck is a real number and each hk is an indicator function. Assume that f is an indicator function. In other words. Now that we know that (4. We can always construct a sequence of nonnegative simple functions fn n = 1 2 : : : such that 0 f1 (x) f2 (x) f3 (x) : : : for every x 2 IR and f (x) = limn!1 fn (x) for every x 2 IR.e. We include a proof of this because it allows us to illustrate the concept of the standard machine explained in Williams’s book in Section 5.5) and “cancel” the d 0 ). Introduction to Probability Theory 39 dIP an equation which is suggested by the notation introduced in (4.CHAPTER 1.12.. i. In IP (A) = ' d 0: A This is true because it is the definition of IP (A). we consider a general nonnegative function f . Step 2.. (4.5) becomes Z = l A (x) for some Borel set A I IR. The standard machine argument proceeds in four steps. i. Step 1.5) holds when f is an indicator function.e. We have already proved that Z We let n ! 1 and use the Monotone Convergence Theorem on both sides of this equality to get Z Z IR fn dIP = Z IR fn ' d 0 for every n: IR f dIP = IR f' d 0: . assume that f simple function. page 5. a linear combination of indicator functions.

although most of the examples we give will be for coin toss space. but you are told that ! 2 B . the conditional probability you assign to B is still the unconditional probability IP (B ).1 Independence of sets Definition 1. IR f 'd . Z ZIR IR f +' d .. f +' d f' d 0 : Z 0. with probability p for H and probability q = 1 .40 Step 4. To avoid trivialities. 1. This discussion is symmetric with respect to A and B . 0 f ' d 0: Subtracting these two equations.5. In the last step. (A \ IP (AjB) = IP IP (B)B) : Whether two sets are independent depends on the probability measure IP . 0 1. we mean that Z IR ¿From Step 3. which can take both positive and negative values. i. if A and B are independent and you know that ! 2 A. we obtain the desired result: Z IR f dIP = = = Z ZIR ZIR R f + dIP . For example. p for T on each toss. the probability that ! 2 A is The sets A and B are independent if and only if this conditional probability is the uncondidtional probability IP (A). knowing that ! 2 B does not change the probability you assign to A.e. Conditional on this information. we consider an integrable function f . suppose we toss a coin twice. Suppose you are not told !. we define and discuss the notion of independence in a general probability space ( F IP ). Then IP fHH g = p2 IP fHT g = IP fTH g = pq IP fTT g = q2: (5. By integrable.1) . IR Z f dIP . and ! is the outcome.15 We say that two sets A 2 F and B IP (A \ B) = IP (A)IP (B): 2 F are independent if Suppose a random experiment is conducted. we assume that 0 < p < 1. Z ZIR IR f + dIP = f dIP = .5 Independence In this section. we have f + dIP <1 Z IR f dIP < 1: . The probability that ! 2 A is IP (A).

In fact. i. IP (B ) = 0:0198. and the probability of one head and one tail is quite small.e. and let IP be given by (5. the probability of one head and one tail. is still 1 . A is the set “H on the first toss” and B is the = fHT g. if you are told that the first toss resulted in H . Introduction to Probability Theory Let A = fHH HT g and B = fHT set “one H and one T .5.1). We say that G and H are independent if every set in G is independent of every set in H. we will find that the product of the probabilties is the probability of the intersection.2 Independence of -algebras Definition 1. conditioned on getting a H on the first toss.-algebras of F . which is the case if and only if p = 1 2. in fact. which is now the probability of a T on the second toss. 1. the probability of one H and one T is the probability of a T on the second toss. if we choose the set fHH the set fHH TH g from H. Let determined by the first toss: G contains the sets G = F1 be the -algebra fHH HT g fTH TT g: Let H be the -albegra determined by the second toss: H contains the sets fHH TH g fHT TT g: These two -algebras are independent. which is 0:99. These sets are independent if and only if 2p 2 q = pq . . IP (A \ B) = IP (A)IP (B) for every A 2 H B 2 G : Example 1. then we have HT g from G and IP fHH HT gIP fHH TH g = (p2 + pq )(p2 + pq ) = p2 IP fHH HT g\ fHH TH g = IP fHH g = p2: No matter which set we choose in G and which set we choose in H. We compute IP (A) = p2 + pq = p IP (B ) = 2pq IP (A)IP (B ) = 2p2q IP (A \ B) = pq: If p = 1 . For example.CHAPTER 1. By far the most likely outcome of the two coin tosses is TT . it becomes very likely that the two tosses result in one head and one tail. the probability of B . In words. If you are told that the coin 2 2 tosses resulted in a head on the first toss. then IP (B ). However. 2 Suppose however that p = 0:01.” Then A \ B 41 TH g.16 Let G and H be sub. is 1 .14 Toss a coin twice.

3 Independence of random variables Definition 1. = fHH HT TH TT g to be 1 1 IP fHH g = 9 IP fHT g = 2 IP fTH g = 1 IP fTT g = 3 9 3 and for every set A . whereas S3 is S2 S2 either u or d. In the case of S2 and S3 just considered. We defined earlier the measure induced by X on IR to be Similarly. the sets fS2 = are indepedent sets. In the probability space of three independent coin tosses.14 illustrates the general principle that when the probability for a sequence of tosses is defined to be the product of the probabilities for the individual tosses of the sequence. LX Y (C ) = IP f(X Y ) 2 C g C IR2: f(X Y ) 2 A Bg = fX 2 Ag \ fY 2 Bg . and we can define the measure induced by the The set C in this last equation is a subset of the plane IR 2 . C could be a “rectangle”. 4 4 9 . as shown by the following example. define IP (A) to be the sum of the probabilities of the elements in A. so the product of the probabilities is 27 . It is also possible to construct probabilities such that the different tosses are not independent. We say that the different tosses are independent when we construct probabilities this way. udS0g = fHTH HTT g and n S3 S2 = u = fHHH HTH THH TTH g o Suppose X and Y are independent random variables.42 Example 1. which has probability 1 . On the other hand. These sets are not 9 1. Then IP ( ) = 1. Definition 1. In this case. i. every set defined in terms of X is independent of every set defined in terms of Y . Note that the sets fH on first tossg = fHH HT g and fH on second tossg = fHH TH g have probabilities IP fHH HT g = 1 and IP fHH TH g = 3 Example 1. the price S 2 of the stock at time 2 is independent of S3 . so IP is a probability measure. for exS2 ample. the intersection of fHH HT g and fHH TH g contains the single element fHH g.17 says that for independent random variables X and Y . then every set depending on a particular toss will be independent of every set depending on a different toss. the measure induced by Y is Now the pair (X pair LX (A) = IP fX 2 Ag A IR: LY (B) = IP fY 2 Bg B IR: Y ) takes values in the plane IR 2. where A IR and B IR. depending on whether the third coin toss is H or T . In particular.15 Define IP for the individual elements of independent.5.e. a set of the form A B .17 We say that two random variables X and Y are independent if the -algebras they generate (X ) and (Y ) are independent. This is because S2 depends on only the first two coin tosses.

1 y ]. Then X and Y are independent variables if and only if the joint density is the product of the marginal densities.1 fX Y ( y ) d : LX (A) = LY (B) = Z ZA B fX (x) dx A IR fY (y) dy B IR: Suppose X and Y have a joint density. Therefore. Since every set in (X ) is independent of every set in (Y ). But f! W (!) 2 Ag = f! : g(X (!)) 2 Ag which is defined in terms of the random variable X . Let g and h be functions from IR to IR.1 fX Y (x ) d fY (y) Z 1 . write (5. Introduction to Probability Theory and X and Y are independent if and only if 43 LX Y (A B) = IP fX 2 Ag \ fY 2 Bg = IP fX 2 AgIP fY 2 B g = LX (A)LY (B ): (5.7 Suppose X and Y are independent random variables. Theorem 5. . X can contain strictly more information than W . P ROOF : Let us denote W = g (X ) and Z a typical set in (W ) is of the form = h(Y ).) In the same way that we just argued that every set in (W ) is also in (X ). then the random variables X and Y have marginal densities defined by A B fX Y (x y) dx dy: fX (x) = These have the properties 1 . and differentiate with respect to both x and y . Then g(X ) and h(Y ) are also independent random variables. this is the case. the joint distribution represented by the measure LX Y factors into the product of the marginal distributions represented by the measures LX and LY .2) is equivalent to independence of X and Y . for independent random variables X and Y . which means that (X ) will contain all the sets in (W ) and others besides. This follows from the fact that (5.1) in terms of densities. we conclude that every set in (W ) is independent of every set in (Z ). this set is in (X ).2) In other words.1 x] and B = (.CHAPTER 1. Take A = (. A joint density for (X Y ) is a nonnegative function f X Y (x y) such that LX Y (A B) = Z Z Z Not every pair of random variables (X Y ) has a joint density. we have that every set in (W ) is also in (X ). but if a pair does. which means that X contains at least as much information as W . for example. if W = X 2. (In general. In fact. we can show that every set in (Z ) is also in (Y ). We must consider sets in (W ) and (Z ).

8 If two random variables X and Y are independent. we define their correlation coefficient (X Y ) = p Cov(X Y ) : Var(X )Var(Y ) For independent random variables. Define Y = XZ . The variance of a random variable X is defined to be Var(X ) = IE X . X and Y were independent. Now use the standard machine to get the result for general functions g and h.18 Let X1 X2 : : : be a sequence of random variables.5. two random variables can have zero correlation and still not be independent. IEX IEY: h i X and Y According to Theorem 5. IP (A1 \ A2 \ An) = IP (A1 )IP (A2 ) IP (An ): 1.16 Let X be a standard normal random variable. IEX )(Y . the correlation coefficient is zero. Unfortunately. so would be X 2 and Y 2 . We show that Y is also a standard normal random variable.44 Definition 1. the covariance is zero. . Consider the following example. I Then the equation we are IP fX 2 Ag \ fY 2 Bg = IP fX 2 AgIP fY 2 Bg which is true because X and Y are independent. IEX ]2: The covariance of two random variables X and Y is defined to be Cov(X Y ) = IE (X . We say that these random variables are independent if for every sequence of sets A1 2 (X1) A2 2 (X2) : : : and for every positive integer n. let Z be independent of X and have the distribution IP fZ = 1g = IP fZ = .8. I P ROOF : Let g (x) = l A (x) and trying to prove becomes h(y) = l B (y ) be indicator functions.1g = 0. but in fact. The last claim is easy to see. but X and Y are not independent. IEY ) = IE XY ] . Example 1.4 Correlation and independence Theorem 5. then IE g (X )h(Y )] = IEg (X ) IEh(Y ) provided all the expectations are defined. X and Y are uncorrelated. If both have positive variances. and if g and h are functions from IR to IR. for independent random variables. If X 2 = Y 2 almost surely.

Indeed. If we are given independent random variables X1 X2 : : : Xn . the same as the best estimate of X based on no information. we shall need only the second assertion of the property: (k) If a random variable X is independent of a -algebra H. then IE X jH] = IEX: The point of this statement is that if X is independent of H. Therefore.e. Cov(X Since X is standard normal. 1995. then the best estimate of X based on the information in H is IEX . Introduction to Probability Theory We next check that Y is standard normal. IEX . IEZ )2 i = IE X + Y . the variance of their sum is the sum of their variances.1g y gIP fZ = 1g + IP f.CHAPTER 1. page 88. IEX )(Y .3) 1.. The property as stated there is taken from Williams’s book.X y g: 2 .yg. then Var(Z ) = IE (Z . if X and Y are independent and Z = X + Y . We now return to property (k) for conditional expectations. For y 45 IP fY Being standard normal. IEY ) + (Y . yg = IP fY = IP fX = IP fX 1 = 2 IP fX 2 IR. what is the distribution of (X We conclude this section with the observation that for independent random variables. then Var(X1 + X2 + + Xn ) = Var(X1) + Var(X2) + + Var(Xn ): (5. IEY ] + Var(Y ) = Var(X ) + Var(Y ): h i This argument extends to any finite number of random variables. we have y and Z = 1g + IP fY y and Z = . IEY )2 = Var(X ) + 2IE X . and we have IP fY yg = IP fX yg. IP fX y g = IP fX which shows that Y is also standard normal. presented in the lecture dated October 19.1g y g + 1 IP f.X y and Z = .5 Independence and conditional expectation. IEY )2 h i = IE (X . both X and Y have expected value zero.X y gIP fZ = . i. Y ) = IE XY ] = IE X 2Z ] = IEX 2 IEZ = 1 0 = 0: Y )? Where in IR2 does the measure LX Y put its mass. IEX )2 + 2(X . .5.1g y and Z = 1g + IP f. IEX ]IE Y .

but here is an argument which makes it plausible.46 To show this equality. n X k=1 Var n Xk = X 2 = 2 : 2 n n k=1 n . and the right hand side is A IP (B) dIP = Z A I l B dIP: Z I I l Al B dIP = Z I l A B dIP = IP (A \ B): \ The partial averaging equation holds because A and B are independent. Define the sequence of averages Then Yn converges to Yn = X1 + X2 + + Xn n = 1 2 : : :: n almost surely as n ! 1. we first recall from (5. each with expected value and variance 2. We must also check the partial averaging property Z If X is an indicator of some set B . In other words. For the first step. which by assumption must be independent of H. We next check that Var(Yn ) ! 0 as n ! 0. we observe first that IEX is H-measurable. then the partial averaging equation we must check is A IEX dIP = Z A X dIP for every A 2 H: Z The left-hand side of this equation is IP (A)IP (B ). Here is the first one. Theorem 5. We are not going to give the proof of this theorem. The partial averaging equation for general X independent of H follows by the standard machine. since it is not random.6 Law of Large Numbers There are two fundamental theorems about sequences of independent random variables. We first check that IEYn = for every n. identically distributed random variables. The argument proceeds in two steps. 1. Therefore.5. the random variables Yn are increasingly tightly distributed around as n ! 1. we have Var(Yn ) ! 0.3) that the variance of the sum of independent random variables is the sum of their variances. Var(Yn ) = As n ! 1. We will use this argument later when developing stochastic calculus. we simply compute 1 IEYn = n IEX1 + IEX2 + 1 + IEXn] = n | + + + } = : ] {z n times For the second step.9 (Law of Large Numbers) Let X1 X2 : : : be a sequence of independent.

pn = n X k=1 2 n = : 1 Z e lim IP fZn 2 Ag = p n 2 A !1 .5. we should divide by n rather than n. the distribution of Z n approaches that of a normal random variable with mean (expected value) zero and variance 2. k=1 Xk .7 Central Limit Theorem The Law of Large Numbers is a bit boring because the limit is nonrandom. This is because the denominator in the definition of Y n is so large that the variance of Yn converges to zero. x2 2 2 dx: .pn + n X Var + (Xn . ) + (X2 . if we again have a sequence of independent. If we want p to prevent this. for every set A IR. around their expected value 0. as measured by their variance. identically distributed random variables. ) 2 As n ! 1. the distributions of all the random variables Z n have the same degree of tightness. In other words. Introduction to Probability Theory 47 1. each with expected value and variance 2 .CHAPTER 1. The Central Limit Theorem asserts that as n ! 1. but now we set then each Zn has expected value zero and Var(Zn ) = ) Zn = (X1 . In particular.

48 .

sequences of tosses of length 3) is = fHHH HHT HTH HTT THH THH THT TTH TTT g: A typical sequence of will be denoted ! . . Thus in the 3-coin-toss example we write for instance. S1 (!) = S1(!1 !2 !3) = S1(! 1) 4 4 S2 (!) = S2(! 1 !2 !3) = S2(!1 !2): 4 4 Each Sk is a random variable defined on the set . and down by a factor of d if it comes out tails (T ).Chapter 2 Conditional Expectation Please see Hull’s book (Section 9. We will see later that Sk is in fact R 49 . the stock price either goes up by a factor of u or down by a factor of d. after k tosses) under the outcome !.1) The collection of all possible outcomes (i. Then F is a -algebra and ( F ) is a measurable space. and ! k will denote the kth element in the sequence ! . 2. At each time step. It will be useful to visualize tossing a coin at each time step.e. Note that Sk (! ) depends only on ! 1 !2 : : : !k . that is.) 2. and say that the stock price moves up by a factor of u if the coin comes out heads (H ). let F = P ( ).e. Note that we are not specifying the probability of heads here. We write Sk (! ) to denote the stock price at “time” k (i. Sk 1 is a function B!F where B is the Borel -algebra on I . Each Sk is an F -measurable function !IR.6.1 A Binomial Model for Stock Price Dynamics Stock prices are assumed to follow this simple binomial model: The initial stock price during the period under study is denoted S 0. More precisely. Consider a sequence of 3 tosses of the coin (See Fig.

the collection F 1 of sets determined by the first toss consists of: .1 (Sets determined by the first k tosses. for each k = 1 2 : : : n. R R For any random variable X defined on a sample space 4 and any y 2 IR.-algebra of F . It is easy to check that F k is a -algebra. Similarly for any subset B of IR. we will use the notation: fX yg = f! 2 X (!) yg: The sets fX < y g fX y g fX = y g etc. In this course we will always deal with subsets of I that belong to B.) We say that a set A is determined by the first k coin tosses if. Recall that the Borel -algebra B is the -algebra generated by the open intervals of I .1 In the 3 coin-toss example. we can decide whether the outcome of all tosses is in A.2 Information Definition 2. we define fX 2 Bg = f! 2 X (!) 2 Bg: 4 Assumption 2. are defined similarly.1 u > d > 0. Note that the random variable Sk is F k -measurable. knowing only the outcome of the first k tosses. 2. Example 2. measurable under a sub.1: A three coin period binomial model.50 ω = Η 3 S2 (HH) =u 2S0 ω2 = Η S (H) = uS0 1 ω1= Η S0 ω1 = Τ S (T) = dS0 1 ω2 = Τ S2 (TT) = d 2S0 ω3 = Τ S3 (TTT) = d 3 S0 ω = Η 3 ω2 = Η ω2 = Τ S2 (HT) = ud S 0 S2 (TH) = ud S 0 ω = Η 3 ω3 = Τ ω3 = Τ S3 (HHT) = u2 d S0 S3 (HTH) = u2 d S0 S3 (THH) = u2 d S0 3 S3 (HHH) = u S 0 S3 (HTT) = d 2 u S0 S3 (THT) = d 2 u S0 S3 (TTH) = d 2 u S0 Figure 2. In general we denote the collection of sets determined by the first k tosses by F k .

then (X ) is given by . we can decide whether or not ! 2 A. . 51 4 AH = fHHH HHT HTH HT T g. and .2 (Sets determined by S2 ) The -algebra generated by S2 consists of the following sets: 1. which we call the -algebra generated by X . 5. If the random variable X takes finitely many different values. AT T = fTTH T TT g = fS2 = d2S0 g AHT AT H = fS2 = udS0g 4. Complements of the above sets. !IR. Another way of saying this is that for every y 2 IR. We say that a set A is determined by the random variable X if. if X is a random variable (X ). knowing only the value X (!) of the random variable. 4 AT H = fTHH THT g. 7. 7. 4 AHT = fHTH HTT g. and denote by (X ).CHAPTER 2. 5. .) Let X be a random variable !IR. The complements of the above sets. The collection F 2 of sets determined by the first two tosses consists of: 1.1 (y ) \ A = . 4 AT = fTHH THT TTH TTT g. 4 AT T = fTTH T TT g. 4. = fS2 (!) 2 IRg. 3. then tion of sets (X ) is generated by the collec- fX 1(X (!))j! 2 g . . 2. Any union of the above sets (including the complements). 4 AHH = fHHH HHT g. The collection of susbets of determined by X is a -algebra. Any union of the above sets. 2.1(y ) A or X . = fS2 (!) 2 g. 4.2 (Information carried by a random variable. AHH = fHHH HHT g = f! 2 S2 (!) = u2S0 g. Conditional Expectation 1. these sets are called the atoms of the -algebra In general. 2. 6. 3. 3. 6. Definition 2. either X . (X ) = fX 1(B ) B 2 Bg: Example 2.

If the value of S2 is known. 8A 4 2 . then even without knowing ! . ! A X (!)IP (! ): 2. We define IE (S1jS2)(TTT ) = IE (S1jS2 )(TTH ) = dS0 : . 4 the coin tosses are independent.) IEX = 4 X ! 2 X (!)IP (! ): 1 0 if ! if ! If A then IA (!) = 4 ( and IE (IAX ) = Z A XdIP = 2A 62 A X 2 We can think of IE (IA X ) as a partial average of X over the set A. i. we need to introduce a probability measure on our coin-toss sample space . We define IE (S1jS2)(HHH ) = IE (S1jS2)(HHT ) = uS0: – If ! = TTT or ! = TTH .3 Conditional Expectation In order to talk about conditional expectation. given S2.52 2.1 An example Let us estimate S1. so that. If we know that S2(! ) = u2 S0 . then S2 (! ) = u2 S0.3. IE (S1jS2) is a random variable Y whose value at ! is defined by Y (!) = IE (S1jS2 = y ) where y = S2 (!). we know that S 1 (! ) = dS0. IP (HHT ) = p2q etc. From elementary probability. we know that S 1 (! ) = uS0. it is a random variable.. p) is the probability of T . If we know that S2(! ) = d2S0 . then even without knowing ! . Denote the estimate by IE (S1jS2). – If ! = HHH or ! = HHT ..3 (Expectation. Let us define p 2 (0 1) is the probability of H . then S2(! ) = d2S0 .e. Properties of IE (S1jS2): IE (S1jS2) should depend on !.g. IP (A) = P! A IP (!). Definition 2. In particular. then the value of IE (S 1jS2 ) should also be known. e. q = (1 .

IE (S1jS2) is random only through dependence on S 2 . then S2(!) = udS0. then we do not know whether S1 = uS0 or S1 = dS0.-algebra of F . and let G be a sub.83. we can write where In other words. Then IE (X jG) is defined to be any random variable Y that satisfies: Y is G -measurable.2 Let ( on ( (a) IE (S1jS2)dIP = Z A S1dIP: Definition of Conditional Expectation Please see Williams. The random variable IE (S1jS2) has two fundamental properties: IE (S1jS2) is (S2)-measurable. We then take a weighted average: 2 A = fHTH HTT THH THT g. Let X be a random variable F IP ).3. Z A 2. Conditional Expectation – If ! 53 udS0. If we know S2(!) = IP (A) = p2 q + pq2 + p2q + pq 2 = 2pq: Furthermore. Z A S1 dIP = p2quS0 + pq2uS0 + p2qdS0 + pq 2dS0 = pq (u + d)S0 For ! 2 A we define R S dIP IE (S1jS2)(!) = A (1A) = 1 (u + d)S0: 2 IP Z A Then IE (S1jS2)dIP = Z A S1dIP: In conclusion. F IP ) be a probability space. For every set A 2 (S2). . We also write 8 > uS0 < g(x) = > 1 (u + d)S0 2 : dS 0 IE (S1jS2)(!) = g (S2(!)) if x = u2 S0 if x = udS0 if x = d2 S0 IE (S1jS2 = x) = g (x) where g is the function defined above. p.CHAPTER 2.

although the dependence on ! is often not shown explicitly. There can be more than one random variable Y satisfying the above properties. conditional expectations always exist. The value of ! is partially but not fully revealed to us. There is always a random variable Y satisfying the above properties (provided that IE jX j < 1). then Y = Y 0 almost surely. but if Y 0 is another one. it is standard notation to write 4 IE (X jY ) = IE (X j (Y )): Here are some useful ways to think about IE (X jG): A random experiment is performed. and thus we cannot compute the exact value of X (!).3. (3. we have the “partial averaging property” Z A Y dIP = Z A XdIP: Uniqueness. i.3 Further discussion of Partial Averaging The partial averaging property is Z A We can rewrite this as IE (X jG)dIP = Z A XdIP 8A 2 G : (3.10 If V is any G -measurable random variable. which is the intersection of all sets in G containing ! . Thus. but not told which element of A it is.2) IE V:IE (X jG)] = IE V:X ]: (3. IP f! 2 Y (!) = Y 0 (!)g = 1: Notation 2.e.e.e. If the -algebra G contains finitely many sets.e. it depends on !.1) IE IA:IE (X jG)] = IE IA:X ]: Note that IA is a G -measurable random variable. i.54 (b) For every set A 2 G . IE X jY ](!) will be the same. we compute an estimate of X (!). Based on what we know about ! . an element ! of is selected. for all ! in this set A. In fact the following holds: Lemma 3... Y .. We then define IE X jY ](!) to be the average (with respect to IP ) value of X over this set A.3) . then provided IE jV:IE (X jG)j < 1. 2. IE X jY ](!) is a function of ! . i.1 For random variables X Existence.. there will be a “smallest” set A in G containing ! . i. Because this estimate depends on the partial information we have about ! . The way ! is partially revealed to us is that we are told it is in A.

3) for each Vn and then pass to the limit. i. Such a V can be written as the limit of an increasing sequence of simple random variables Vn . then the best estimate of X based on G is X itself.3. using the Monotone Convergence Theorem (See Williams).3. unless IP (A) = 0. Based on this lemma. IP (A) = 0. first use (3. Therefore. 56).4) Please see Willams p. V .3) holds for V + and V . Conditional Expectation 55 Proof: To see this. . then IE (X jG) = X: Proof: The partial averaging property holds trivially when Y is replaced by X . it must hold for V as well. Next consider the case that V is a nonnegative G -measurable random variable.CHAPTER 2. we can replace the second condition in the definition of a conditional expectation (Section 2. 88. to obtain (3. V is of the form V = Pn=1 ck IAK . And since X is G -measurable.3) for V . the general G measurable random variable V can be written as the difference of two nonnegative random-variables V = V + . where each Ak is in G and k each ck is constant.2) by: (b’) For every G -measurable random-variable V . then (c) (Linearity) (d) (Positivity) If X IE (X jG) 0: Proof: Take A = f! 2 R (X jG)(! ) < 0g. (a) (b) IE (IE (X jG)) = IE (X ): Proof: Just take A in the partial averaging property to be . Williams calls this argument the “standard machine” (p. . If the information content of G is sufficient to determine X .e. X satisfies the requirement (a) of a conditional expectation as well. and the left-hand side is strictly negative. we have IE V:IE (X jG)] = IE V:X ]: 2. IE R Partial averaging implies A IE (X jG)dIP = A XdIP . The right-hand side is greater than or equal to zero. Proof sketches of some of the properties are provided below. and since (3. If X is G -measurable.3) when V is a simple G -measurable random variable.2) and linearity of expectations to prove (3. Finally..4 Properties of Conditional Expectation (3. we write (3. The conditional expectation of X is thus an unbiased estimator of the random variable X . but is not necessarily simple. This set is in G since IE (X jG) is G -measurable. IE (a1X1 + a2X2jG) = a1IE (X1jG) + a2IE (X2jG): 0 almost surely.

Y also satisfies property (b).56 (h) (Jensen’s Inequality) If : R!R is convex and IE j (X )j < 1. . then IE (IE (X jG)jH) = IE (X jH): -algebra of G means that G contains more information than H. the G -measurable random variable Z acts like a constant. then we get the same result as if we had estimated X directly based on the information in H. as we can check below: Z A Y dIP = IE (IA:Y ) = IE IAZIE (X jG)] = IE IAZ:X ] ((b’) with V = IA Z Z = ZXdIP: A (k) (Role of Independence) If H is independent of ( (X ) G). Then Y satisfies (a) (a product of G -measurable random variables is G-measurable). R R (b) A Y dIP = A ZXdIP When conditioning on G . then IE ( (X )jG) Recall the usual Jensen’s Inequality: IE (IE (X jG)): (IE (X )): (X ) (i) (Tower Property) If H is a sub.-algebra of G . If we estimate X based on the information in G . then IE (X j (G H)) = IE (X jG): In particular. and then estimate the estimator based on the smaller amount of information in H. 8A 2 G . Take Y = Z:IE (X jG). A random variable Y is IE (ZX jG) if and only if (a) Y is G -measurable. then nothing is gained by including the information content of H in the estimation of X . then H is a sub- IE (ZX jG) = Z:IE (X jG): Proof: Let Z be a G -measurable random variable. then IE (X jH) = IE (X ): If H is independent of X and G . if X is independent of H. (j) (Taking out what is known) If Z is G -measurable.

Z AH S2 dIP = p2u2 S0 + pqudS0: Therefore.5 Examples from the Binomial Model Recall that F 1 = f AH AT g. IE (S2jF 1 )(!) = (pu + qd)S1(!) 8! 2 AT : IE (S2jF 1)(!) = (pu + qd)S1(!) 8! 2 : Thus in both cases we have A similar argument one time step later shows that IE (S3jF 2)(!) = (pu + qd)S2(! ): We leave the verification of this equality as an exercise. from the previous equations we have IE IE (S3jF 2)jF 1] = IE (pu + qd)S2jF 2] = (pu + qd)IE (S2jF 1) = (pu + qd)2S1: This final expression is IE (S 3jF 1). We can also write IE (S2jF 1 )(!) = pu2 S0 + qudS0 8! 2 AH : IE (S2jF 1)(! ) = pu2 S0 + qudS0 = (pu + qd)uS0 = (pu + qd)S1(!) 8! 2 AH Similarly. (linearity) . Z Z We compute AH AT IE (S2jF 1)dIP = IE (S2jF 1)dIP = Z Z AH AT S2 dIP S2dIP: Z AH IE (S2jF 1 )dIP = IP (AH ):IE (S2jF 1 )(!) = pIE (S2jF 1)(! ) 8! 2 AH : On the other hand.CHAPTER 2. Notice that IE (S2jF 1 ) must be constant on AH and AT . Now since IE (S2jF 1 ) must satisfy the partial averaging property. Conditional Expectation 57 2. for instance. We can verify the Tower Property.3.

i. then you know the value of Mk . For each k. i. 2.4 Martingales The ingredients are: A probability space ( A sequence of F . Martingales tend to go neither up nor down. Example 2. A supermartingale tends to go down. This -algebra contains no information. Such a sequence of Conditions for a martingale: F IP ). This is called a stochastic process.58 2. If (pu + qd) = 1 then fSk If (pu + qd) 1 then fSk If (pu + qd) 1 then fSk F k k = 0 1 2 3g is a martingale. -algebras F 0 F 1 : : : F n . We say that the process fMk g is adapted to the filtration fF k g.e. and any F 0 -measurable random variable must be constant (nonrandom). we set F 0 = f g. If you know the information in F k . IE(S1 jF 0 ) is that constant which satisfies the averaging property Z IE(S1 jF 0 )dIP = Z S1 dIP: The right hand side is IES1 = (pu + qd)S0 . a submartingale tends to go up. with the property that F 0 F 1 : : : F n -algebras is called a filtration. the “trivial -algebra”.) For k = 1 2 we already showed that IE(Sk+1 jF k ) = (pu + qd)Sk : For k = 0.3 (Example from the binomial model. A sequence of random variables M0 M1 : : : Mn.e. IE (Mk+1 jF k ) = Mk . Each Mk is F k -measurable. 1. by definition. F k k = 0 1 2 3g is a submartingale. F k k = 0 1 2 3g is a supermartingale. . Therefore. and so we have IE(S1 jF 0) = (pu + qd)S0 : In conclusion. IE (M k+1jF k ) Mk . the second condition above is replaced by IE (M k+1 jF k ) Mk .

1 Binomial Pricing Return to the binomial pricing model Please see: Cox. S0 = 50. Cash outlay is . Borrow $40.$40: 59 . The value of the call at time 1 is V1(!) = (S1 (!) . (In this and all examples.$60: 2. 50)+ = 50 0 if ! 1 if ! 1 =H =T Suppose the option sells for $20 at time 0. the interest rate quoted is per unit time. Cash outlay is . and Cox and Rubinstein (1985). 229–263. and the stock prices S0 S1 : : : are indexed by the same time periods). Prentice-Hall. 3. Buy 2 shares of stock for $50 each. Example 3. Sell 3 options for $20 each. We know that S1 (!) = 100 25 if ! 1 if ! 1 =H =T Find the value at time zero of a call option to buy one share of stock at time 1 for $50 (i. Let us construct a portfolio: 1. Ross and Rubinstein.1 (Pricing a Call Option) Suppose u = 2 d = 0:5 r = 25%(interest rate). the strike price is $50).e. Options Markets. 7(1979). J. Financial Economics.Chapter 3 Arbitrage Pricing 3. Cash outlay is $100.

. For this portfolio. .. !1 = H $150 . 0 shares of stock at time 0.$200 $50 $0 !1 = T $0 .. the cash outlay at time 1 is: Pay off option Sell stock Pay off debt $0 The arbitrage pricing theory (APT) value of the option at time 0 is V0 = 20. this is why it does not make sense to use some stock unrelated to the derivative security in valuing it.. ( 0 is also to be determined later) Invest V0 . where V1 is an F 1 -measurable random variable. i. Sell the security for V0 at time 0. his/her trading does not move the market. . (1 + r)S0): We want to choose V0 and 0 so that X1 = V1 regardless of whether the stock goes up or down. Assumptions underlying APT: Unlimited short selling of stock. 0S0 in the money market. at risk-free interest rate r. 0S0 ) (1 + r)V0 + 0 (S1 . (V0 is to be determined later).e. No transaction costs.$50 $50 3..... Buy Then wealth at time 1 is 0S0 might be X1 = 4 = 0 S1 + (1 + r)(V0 . Unlimited borrowing. Important Observation: The APT value of the option does not depend on the probabilities of H and T .60 This portfolio thus requires no initial investment.2 General one-step APT Suppose a derivative security pays off the amount V 1 at time 1. negative). . (This measurability condition is important.. (V0 . Agent is a “small investor”. at least in the straightforward method described below).

(1 + r)S0) V1 = V1 (H ) . 1 . (V (H ) . r)S0 u. V (T ))(u . . Thus the price of the call at time 0 is given by r 1 p = 1 + . d q = u .1): (1 + r)V0 = V1 (H ) . we have 0 < p < 1 and q = 1 .4) 3. Since p + q = 1. 0 (S1(H ) . 1 . Construct a probability measure by the formula 4 f IP (! 1 !2 : : : !n ) = p#fj !j =H g q #fj !j =T g ~ ~ f IP on f f f IP is called the risk-neutral probability measure. d)V (H ) .3 Risk-Neutral Probability Measure Let be the set of possible outcomes from n coin tosses. V1(T ) : S1(H ) .d r 1 = 1 + . We now also assume d be an arbitrage opportunity). )S (T ) (u .3) for V1(H ) . (1 + r)S0) = V1 (H ) 0 (S1(T ) .e. r : ~ ~ u d u d 1 ~ ~ V0 = 1 + r pV1(H ) + qV1(T )]: q are like ~ (2.1) (2. d V1(H ) + u . Arbitrage Pricing 61 The last condition above can be expressed by two equations (which is fortunate since there are two unknowns): (1 + r)V0 + (1 + r)V0 + 0(S1 (H ) .4 says f 1 V0 = IE 1 + r V1 : . Subtracting the second equation above from the first gives 0= Plug the formula (2. i. p ~ ~ ~ ~ ~ ~ ~ ~ probabilities.CHAPTER 3. . (1 + r)S0) = V1 (T ) (2. . Define 4 4 1 + r u (otherwise there would Then p > 0 and q > 0. We will return to this later. Equation 2.3) 0 into (2. r)] 1 1 1 u. Thus. r V1(T ): u d u d We have already assumed u > d > 0. . We denote by IE the expectation under IP . .d 0 = 1 (u .. when Sk is known for a given ! .2) Note that this is where we use the fact that the derivative security value V k is a function of Sk .. p. V1((H ) . Vk is known (and therefore non-random) at that ! as well. S1(T ) (2.

where . . which need not be 0. . d)(1 + r) Sk u. k Sk+1 + (1 + r)(Xk . d .62 f P Theorem 3. .12 Under I . d) + .1) (3. r) S k u.d + du = (1 + r) (k+1) u + ur . d .3. (No insider trading).2) f P Theorem 3. ud . the discounted stock price process f(1+ r) . Each k is F k -measurable. . . .1 Portfolio Process The portfolio process is =( 0 1 ::: n 1).k Xk is a martingale.d kS : = (1 + r) k . (k+1) Sk+1 jF k ] (1 + r) (k+1)(~u + qd)Sk p ~ (1 + r) (k+1) u(1 + r . Proof: We have F k gn=0 k (1 + r) (k+1)Xk+1 = (1 + r) k Xk + k (1 + r) (k+1) Sk+1 . the discounted self-financing portfolio process value f(1 + r) . k is the number of shares of stock held between times k and k + 1.3.d u. . (1 + r)Sk ): (3.k Sk Proof: F k gn=0 is a martingale. Define recursively Xk+1 = = Then each Xk is F k -measurable. (1 + r) k Sk : . 3. . k f IE (1 + r) = = d(u .2 Self-financing Value of a Portfolio Process Start with nonrandom initial wealth X 0 . 1 . dr Sk u = (1 + r) (k+1) (u .11 Under I . k Sk ) (1 + r)Xk + k (Sk+1 . 3.

.(1 + r) k k Sk (property (b)) = (1 + r) k Xk (Theorem 3. for k = 0 1 : : : m. satisfies the f IE Mk+1 jF k ] = Mk f ff IE Mm jF m 2 ] = IE IE Mm jF m 1 ]jF m 2 ] f = IE Mm 1 jF m 2 ] = Mm 2 : . 2. .1) Proof: We first observe that if martingale property for each k we use the tower property to write = 0 1 : : : m . m is less than or equal to n.. . Arbitrage Pricing Therefore. the number of periods/coin-tosses in the model). .1 () A simple European derivative security with expiration time m is an F m -measurable random variable Vm .IE (1 + r) k k Sk jF k ] = (1 + r) k Xk (requirement (b) of conditional exp. Definition 3. 1. the APT value at time k of V m is f Vk = (1 + r)k IE (1 + r) m VmjF k ]: 4 . 1.11) . then we also have f IE Mm jF k ] = Mk k = 0 1 : : : m . .e.CHAPTER 3.2) follows directly from the martingale property. 63 f IE (1 + r) (k+1)Xk+1 jF k ] f = IE (1 + r) k Xk jF k ] f +IE (1 + r) (k+1) k Sk+1 jF k ] f . . . . . (Here. the equation (4. Theorem 4. then for each k = 0 1 : : : m. we call Xk the APT value at time k of V m . . 1: (4.) f + k IE (1 + r) (k+1) Sk+1 jF k ] (taking out what is known) .2) When k = m .1 ) such that the self-financing value process X0 X1 : : : Xm given by (3.12) If a simple European security Vm is hedgeable.2 () A simple European derivative security Vm is said to be hedgeable if there exists a constant X0 and a portfolio process = ( 0 : : : m.4 Simple European Derivative Securities Definition 3. fMk F k k = 0 1 : : : mg is a martingale. . 3. (4. For k = m . i.13 (Corollary to Theorem 3.2) satisfies Xm(!) = Vm(!) 8! 2 : In this case. . .

Theorem 3. Vk+1 (!1 :: :: :: !k T ) : !k H ) . . the self-financing value of the portfolio process 0 1 : : : m..1) and (5. and so for each k. . the model is said to be complete. f Xk = (1 + r)k IE (1 + r) m VmjF k ]: .1) k (!1 f E Starting with initial wealth V 0 = I (1 + r).3) holds for some value of k. f f (1 + r) k Xk = IE (1 + r) m Xm jF k ] = IE (1 + r) m Vm jF k ]: . the model is called incomplete.12 says that X0 (1 + r) 1X1 : : : (1 + r) m Xm is a martingale. Proof: Let V0 : : : Vm. and set f Vk (!1 : : : !k ) = (1 + r)k IE (1 + r) m Vm jF k ](!1 : : : !k ) . then there is a portfolio process whose selffinancing value process X0 X1 : : : Xm satisfies Xm = Vm . 3.e. Theorem 5.3) holds by definition of X 0. .64 We can continue by induction to obtain (4. Assume that (5.5 The Binomial Model is Complete Can a simple European derivative security always be hedged? It depends on the model. For k = 0. Sk+1 (!1 !k T ) k+1 (!1 (5.2: V : : : !k ) = Sk+1(!1 :: :: :: !k H ) . let V m be a simple European derivative security.2) Xk+1 = We need to show that k Sk+1 + (1 + r)(Xk .1 by the recursive formula 3.1 be defined by (5. for each fixed (!1 : : : !k ). (5. we have Xk (!1 : : : !k ) = Vk (!1 : : : !k ): . By definition. In particular.1 and 0 : : : m. i. (5. Xk is the APT value at time k of Vm .2). If the answer is “yes”.14 The binomial model is complete.1 is the process V0 V1 : : : Vm. Therefore. Set X0 = V0 and define the self-financing value of the portfolio process 0 : : : m.3) We proceed by induction.m Vm ]. If the answer is “no”. If the simple European security Vm is hedgeable. k Sk ): Xk = Vk 8k 2 f0 1 : : : mg: (5. .2).

Vk+1 (T )) q + pVk+1 (H ) + q Vk+1 (T ) ~ ~ ~ = Vk+1 (H ): . Arbitrage Pricing We need to show that 65 Xk+1 (!1 : : : !k H ) = Vk+1(!1 : : : !k H ) Xk+1(!1 : : : !k T ) = Vk+1(!1 : : : !k T ): We prove the first equality.d = (Vk+1 (H ) . f Vk (!1 : : : !k ) = IE (1 + r) 1 Vk+1jF k ](!1 : : : !k ) 1 p ~ = 1 + r (~Vk+1 (!1 : : : !k H ) + q Vk+1 (!1 : : : !k T )) : 1 p ~ Vk = 1 + r (~Vk+1 (H ) + qVk+1(T )) : Since (!1 : : : !k ) will be fixed for the rest of the proof. V = Sk+1 (H ) . 1 . P k . (1 + r)Sk ) k k +~Vk+1 (H ) + q Vk+1 (T ) p ~ = (Vk+1 (H ) . (k+1) Vk+1 jF k ] f In other words.k Vk gn=0 is a martingale under I . Sk+1 (T ) (Sk+1 (H ) . dS+1 (T ) (uSk . f(1 + r). (1 + r)Sk ) + (1 + r)Vk V (H ) . Vk = Vk+1uS ) . (1 + r)Sk ) k+1 k+1 (T ) +~Vk+1 (H ) + q Vk+1 (T ) p ~ (H . k Sk ) = k (Sk+1 (H ) . In particular. .CHAPTER 3. . the second can be shown similarly. = = = ff IE IE (1 + r) m VmjF k+1 ]jF k ] f IE (1 + r) mVm jF k ] (1 + r) k Vk . Vk+1 (T )) u . we simplify notation by suppressing these symbols. Note first that f IE (1 + r) . we write the last equation as We compute Xk+1 (H ) = k Sk+1 (H ) + (1 + r)(Xk . For example. r + pVk+1 (H ) + qVk+1 (T ) ~ ~ u.

66 .

and the value and portfolio processes are given by: f Vk = (1 + r)k IE (1 + r) m Vm jF k ] k = 0 1 : : : m . Vk+1(!1 : : : !k T ) k (!1 : : : !k ) = S (! : : : ! H ) . we can now compute: V . = S1(H) . with payoff given by (See Fig.d = 0:5 q = 1 . 1: k+1 1 k k+1 1 k .1 Binomial Model Pricing and Hedging Recall that Vm is the given simple European derivative security. p = ~ ur d ~ ~ Consider a simple European derivative security with expiration 2. V1 (T ) = 0:93 (H) S (T ) 1 1 2 2 1 V (H) = S2 (HH) . V2 (HT) = 0:67 (HH) . S (! : : : ! T ) k = 0 1 : : : m . we have: Using these values. 4..1): 0:5: V2 = 0max2(Sk . 5)+ : k Notice that V2 (HH) = 11 V2 (HT) = 3 6= V2 (TH) = 0 V2 (T T ) = 0: 1 ~ V1 (H) = 1 + r pV2 (HH) + qV2 (HT)] = 4 0:5 11 + 0:5 3] = 5:60 ~ 5 V1 (T) = 4 0:5 0 + 0:5 0] = 0 5 4 0:5 5:60 + 0:5 0] = 2:24: V0 = 5 0 The payoff is thus “path dependent”.1 (Lookback Option) u = 2 d = 0:5 r = 0:25 S0 = 4 p = 1+. Working backward in time.Chapter 4 The Markov Property 4. 1: Vk+1 (!1 : : : !k H ) . Example 4. S (HT) 67 .

2 (European Call) Let u = 2 with expiration time 2 and payoff function ~ ~ d = 1 r = 1 S0 = 4 p = q = 2 4 V2 = (S2 .1: Stock price underlying the lookback option. and consider a European call Note that V2 (HH) = 11 V2 (HT) = V2 (TH) = 0 V2 (T T ) = 0 1 V1 (H) = 4 2 :11 + 1 :0] = 4:40 2 5 41 V1(T) = 5 2 :0 + 1 :0] = 0 2 41 V0 = 5 2 4:40 + 1 0] = 1:76: 2 Define vk (x) to be the value of the call at time k when Sk = x. 0S0 ) = 0:01 V1 (T ) = 0 X1 (HH) = 1(H)S1 (HH) + (1 + r)(X1 (H) . 0S0 ) = 5:59 V1 (H) = 5:60 X1 (T ) = 0S1 (T) + (1 + r)(X0 . 5)+ v1(x) = 4 1 v2 (2x) + 1 v2 (x=2)] 2 52 v0(x) = 4 1 v1 (2x) + 1 v1 (x=2)]: 2 52 .68 S2 (HH) = 16 S (H) = 8 1 S2 (HT) = 4 S =4 0 S1 (T) = 2 S2 (TH) = 4 S2 (TT) = 1 Figure 4. 5)+ : 1 2. we can check that X1 (H) = 0S1 (H) + (1 + r)(X0 . S (TT) 2 2 Working forward in time. 1(H)S1 (H)) = 11:01 V1 (HH) = 11 etc. Example 4. 1 V (T) = S2 (TH) . V2 (TT ) = 0: (TH) . Then v2 (x) = (x .

heads – so far) and hence only 67 possible option values.2 has three possible values v 2 (16) v2(4) v2(1).. 3. More specifically. Simulation. We have.CHAPTER 4. Then k (x) = vk+1 (2x) . V2(TT ). For period k. x=2 4. then in period 66 there would be 67 possible stock price values (since the final price depends only on the number of up-ticks of the stock price – i. then n = 66 and 266 7 1019. . We could store the value of Vn (! ). 69 Let v2(16) = 11 v2 (4) = 0 v2 (1) = 0 1 v1(8) = 4 2 :11 + 1 :0] = 4:40 2 5 1 v1(2) = 4 1 :0 + 2 :0] = 0 52 1 v0 = 4 2 4:40 + 1 0] = 1:76: 2 5 k (x) be the number of shares in the hedging portfolio at time k when S k = x. rather than four possible values V 2(HH ) V2(HT ) V2(TH ) If there were 66 periods. There are three possible ways to deal with this problem: 1. we must solve 2k 1 ~ Vk (!1 : : : !k ) = 1 + r pVk+1(!1 : : : !k H ) + qVk+1 (!1 : : : !k T )]: ~ For example. and so we could compute V0 by simulation. Look for Markov structure. we could simulate n coin tosses ! = (!1 : : : !n ) under the risk-neutral probability measure. rather than 2 66 7 1019. for example. equations of the form has 2 n elements. The Markov Property In particular. We could repeat this several times and take the average value of Vn as an f EV approximation to I n . We’ll get to that.e.2 Computational Issues For a model with n periods (coin tosses). that f V0 = (1 + r) n IEVn . 2. Then we can use calculus and partial differential equations. Approximate a many-period model by a continuous-time model. In period 2. vk+1 (x=2) k = 0 1: 2x . a three-month option has 66 trading days. If each day is taken to be one period. Example 4. the option in Example 4.2 has this.

70 4.e. and in (c) we assume this condition only for functions h of the form h(x) = e ux . 4. we have IE euXk+1 F k = IE euXk+1 Xk : (If we fix u and define h(x) = eux . Let (The Markov Property). A main result in the theory of Laplace transforms is that if the equation holds for every h of this special form.1 is a function B!B.1 Different ways to write the Markov property (a) (Agreement of distributions). then the equations in (b) and (c) are the same. then it holds for every h. For every A 2 B = B(IR). For every (Borel-measurable) function h : IR!IR for which IE jh(Xk+1 )j < 1.1.e. This process is said to be Markov if: k The stochastic process fXk g is adapted to the filtration fF k g. we have 4 IP (Xk+1 2 AjF k ) = IE IA(Xk+1)jF k ] = IE IA(Xk+1 )jXk ] = IP Xk+1 2 AjXk ]: (b) (Agreement of expectations of all functions). 1.1 () Let ( F P) be a probability space. we have where i = 1.3. we only consider subsets A of I that are in B and functions g : IR!IR such that g . we have IE h(Xk+1 )jF k ] = IE h(Xk+1)jXk]: (c) (Agreement of Laplace transforms. i.. i. However in (b) we have a condition which holds for every function h. (c) implies (b). and F.. (Since jeiuxj = j cos x +sin xj 1 we don’t need to assume that IE jeiuxj < IE eiuXk+1 jF k = IE eiuXk+1 jXk h i h i . Let fF k gn=0 be a filtration under k fXk gn=0 be a stochastic process on ( F P).) p.3 Markov Processes Technical condition always present: We consider only functions on I and subsets of I which are R R R Borel-measurable.) (d) (Agreement of characteristic functions) For every u 2 IR. the distribution of Xk+1 conditioned on F k is the same as the distribution of X k+1 conditioned on X k . For each k = 0 1 : : : n . Definition 4.) For every u 2 IR for which IEeuXk+1 < 1.

we could just as well write g (Xk ) for some function g . Remark. . The Markov property as stated in (a)-(d) involves the process at a “current” time k and one future time k + 1. form (a) of the Markov property can be restated as: For every A 2 B.1 In every case of the Markov properties where IE : : : jXk ] appears. we have IE euk+1 Xk+1+:::+uk+j Xk+j jF k ] = IE euk+1 Xk+1 +:::+uk+j Xk+j jXk ]: (D) For every u = (uk+1 : : : uk+j ) 2 IRj we have IE ei(uk+1Xk+1 +:::+uk+j Xk+j )jF k ] = IE ei(uk+1Xk+1 +:::+uk+j Xk+j )jXk ]: Once again. we have IP (Xk+1 2 AjF k ) = g (Xk ) where g is a function that depends on the set A. IP (Xk+1 : : : Xk+j ) 2 AjF k ] = IP (Xk+1 : : : Xk+j ) 2 AjXk ]: (B) For every function h : IRj !IR for which IE jh(Xk+1 : : : Xk+j )j < 1.CHAPTER 4. For example. Conditions (a)-(d) are equivalent. (A) For every Ak+1 IR : : : Ak+j IR. The Markov Property 71 Remark 4. Conditions (a)-(d) are also equivalent to conditions involving the process at time k and multiple future times. where the function g depends on the random variable represented by : : : in this expression. we have IE h(Xk+1 : : : Xk+j )jF k ] = IE h(Xk+1 : : : Xk+j )jXk ]: (C) For every u = (uk+1 : : : uk+j ) 2 IRj for which IE jeuk+1 Xk+1 +:::+uk+j Xk+j j < 1. Let j be a positive integer. IP Xk+1 2 Ak+1 : : : Xk+j 2 Ak+j jF k ] = IP Xk+1 2 Ak+1 : : : Xk+j 2 Ak+j jXk ]: (A’) For every A 2 IRj . Consequences of the Markov property. We write these apparently stronger but actually equivalent conditions below. every expression of the form IE (: : : jXk ) can also be written as g (Xk ). All these Markov properties have analogues for vector-valued processes.

to obtain (X k ). If we want to estimate the distribution of S k+1 based on the information in F k . e IE Sk+1 jF k ] = (~u + qd)Sk = (1 + r)Sk p ~ (3. the only relevant piece of information is the value of Sk . Consider a model with 66 periods and a simple European derivative security whose payoff at time 66 is 1 V66 = 3 (S64 + S65 + S66 ): . the Markov property requires that for any function h. and (3.1) (Markov property. For example.1)).2).2) is a function of Sk . We will formally prove this later. (with j = 2 in (A)) Assume (b).) (Taking out what is known) (Tower property) (Definition of conditional probability) Now take conditional expectation on both sides of the above equation. IP Xk+1 2 Ak+1 Xk+2 2 Ak+2 jF k ] = IE IAk+1 (Xk+1 )IAk+2 (Xk+2 )jF k ] = IE IE IAk+1 (Xk+1 )IAk+2 (Xk+2 )jF k+1 ]jF k ] = IE IAk+1 (Xk+1 ):IE IAk+2 (Xk+2)jF k+1 ]jF k ] = IE IAk+1 (Xk+1 ):IE IAk+2 (Xk+2)jXk+1 ]jF k ] = IE IAk+1 (Xk+1 ):g (Xk+1)jF k ] = IE IAk+1 (Xk+1 ):g (Xk+1)jXk ] (Markov property. form (b). Then (a) also holds (take h = IA ). form (a). and this is property (A) with j Example 4.2) is the case of h(x) = x. are equal to the RHS of (3.) (Remark 4. Equation (3. Note however that form (b) of the Markov property is stronger then (3. they are equal to each other.3 It is intuitively clear that the stock price process in the binomial model is a Markov process.72 Proof that (b) Consider =) (A).1) IP Xk+1 2 Ak+1 Xk+2 2 Ak+2jXk ] = IE IAk+1 (Xk+1 ):g (Xk+1)jXk]: Since both and IP Xk+1 2 Ak+1 Xk+2 2 Ak+2 jF k ] IP Xk+1 2 Ak+1 Xk+2 2 Ak+2 jXk ] = 2. e IE h(Sk+1)jF k ] is a function of Sk . conditioned on use the tower property on the left.

Assume . the -algebra generated Example 4.. and let G and H be sub. (We are using form (B) of the Markov property here).algebras of F . one of the sets in (S2 ) is f! S2 (!) = u2S0 g = fHH g.4 Showing that a process is Markov Definition 4. For example.66V66 jF 50 ] e = (1 + r).2 (Independence) Let ( F P) be a probability space.e.15 (Independence Lemma) Let X and Y be random variables on a probability space ( F P). We say that G and H are independent if for every A 2 G and B 2 H. (X ). if we take A = fHH HT g from F1 and B = fHH T H g from H. In other words. i.16IE V66jS50] because the stock price process is Markov. which we can determine with a bit of work. the F50-measurable random variable V50 can be written as V50(!1 : : : !50) = g(S50 (! 1 : : : !50)) for some function g.. then IP(A \ B) = IP(HH) = p2 . then IP (A \ B) = IP(HH) = p2 and IP (A)IP(B) = (p2 + pq)(p2 + pq) = p2 (p + q)2 = p2: Note that F 1 and S2 are not independent (unless p = 1 or p = 0). is independent of G .4 Consider the two-period binomial model. Let G be a sub.e. 4. i. we have IP (A \ B) = IP (A)IP (B): We say that a random variable X is independent of a -algebra G if by X . The Markov Property The value of this security at time 50 is 73 e V50 = (1 + r)50IE (1 + r).-algebra of F . H contains the four sets : Then F 1 and H are independent.CHAPTER 4. but IP(A)IP(B) = (p2 + pq)p2 = p3(p + q) = p3 : The following lemma will be very useful in showing that a process is Markov: Lemma 4. For example. Recall that F 1 is the -algebra of sets determined by the first toss. If we take A = fHH HT g from F 1 and B = fHH g from (S2 ). F 1 contains the four sets 4 4 AH = fHH HT g AT = fTH TT g fHH T H g fHT T T g : Let H be the -algebra of sets determined by the second toss.

IE h(Sk+1 )jF k ] as a function of Sk . Define the running maximum of the stock price to be Consider a simple European derivative security with payoff at time n of v n (Sn Examples: Mk = 1maxk Sj : j 4 Mn ) . Let h 4 be any function and set f(x y) = h(xy).5 (Showing the stock price process is Markov) Consider an n-period binomial model. but we have also obtained a formula for 4. Then 4 g(y) = IEf(X y) = IEh(Xy) = ph(uy) + qh(dy): time k and define X The Independence Lemma asserts that IE h(Sk+1 )jF k ] = = = = k IE h SS+1 :Sk jF k ] k IE f(X Y )jG ] g(Y ) ph(uSk ) + qh(dSk ): This shows the stock price is Markov. and define g (y) = IEf (X y): 4 Then IE f (X Y )jG] = g (Y ): Remark.1.5 Application to Exotic Options Consider an n-period binomial model. so that Y is G -measurable. This is a special case of Remark 4. Example 4. Define Y = Sk . we obtain IE h(Sk+1 )jSk ] = ph(uSk ) + qh(dSk ): Thus IE h(Sk+1 )jF k ] and IE h(Sk+1 )jXk ] are equal and form (b) of the Markov property is proved. X is independent of G . In this lemma and the following discussion. . Let f (x y ) be a function of two variables. Fix a 4 k 4 = SS+1 and G = F k . Y is G -measurable.74 X is independent of G . Since X k 4 depends only on the (k + 1)st toss. Not only have we shown that the stock price process is Markov. Indeed. if we condition both sides of the above equation on (S k ) and use the tower property on the left and the fact that the right hand side is (S k )-measurable. capital letters denote random variables and lower case letters denote nonrandom variables. Then X = u if !k+1 = H and X = d if !k+1 = T .

Since (1 + r). Let Z f IP (Z = u) = p IP (Z = d) = q ~ f ~ and Z is independent of F k . Continuing with the exotic option of the previous Lemma. We have Mk+1 = Mk _ Sk+1 4 where _ indicates the maximum of two quantities. . P Proof: Fix k.. K )+ (Lookback option). Let V k denote the value of the derivative f security at time k. Lemma 5. Let h(x k = SS+1 .k Vk is a martingale under I .. so k y ) be a function of two variables.CHAPTER 4. we have Vn = vn (Sn Mn): Stepping back one step. . 1: At the final time. 1 Mn 1)] : . The Markov Property 75 vn(Sn Mn ) = (Mn . we have proved the Markov property (form (b)) for this two-dimensional process. . although that does not matter). . (Here we are working under k the risk-neutral measure I .16 The two-dimensional process f(S k Mk )gn=0 is Markov. K )+ (Knock-in Barrier option). we have P 1 f Vk = 1 + r IE Vk+1jF k ] k = 0 1 : : : n . . We have h(Sk+1 Mk+1 ) = h(Sk+1 Mk _ Sk+1 ) = h(ZSk Mk _ (ZSk )): Define f g (x y) = IEh(Zx y _ (Zx)) = ph(ux y _ (ux)) + q h(dx y _ (dx)): ~ ~ 4 The Independence Lemma implies f IE h(Sk+1 Mk+1 )jF k ] = g(Sk Mk ) = ph(uSk Mk _ (uSk )) + qh(dSk Mk ) ~ ~ the second equality being a consequence of the fact that M k ^ dSk = Mk . vn(Sn Mn ) = IMn B (Sn . we can compute Vn . Since the RHS is a function of (Sk Mk ). 1 = 1 IE v (S M )jF ] f 1+r n n n n 1 1 ~ = 1 + r pvn (uSn 1 uSn 1 _ Mn 1 ) + q vn (dSn ~ .

d))S. which in this case is k _ Mk = vk+1 (uSk (uSk ) u . 1 Mn 1): . . Since this is a simple European option.76 This leads us to define 1 ~ vn 1 (x y) = 1 + r pvn (ux ux _ y ) + qvn(dx y)] ~ 4 . 1 ~ ~ vk (x y) = 1 + r pvk+1 (ux ux _ y ) + q vk+1(dx y ) and the value of the option at time k is v k (Sk Mk ). . the hedging portfolio is given by the usual formula. so that The general algorithm is Vn 1 = vn 1 (Sn . vk+1 (dSk Mk ) ( k .

Again a function g is specified.Chapter 5 Stopping Times and American Options 5. Let Vn = g (Sn ) be the payoff of a derivative security. the hedging portfolio should create a wealth process which satisfies Xk g (Sk) 8k almost surely. This is because the value of the derivative security at time k is at least g (S k ). Consider the Binomial model with n periods.1 American Pricing Let us first review the European pricing formula in a Markov model. d)Sk Now consider an American option. 1: k= (u . In any period k. vk+1 (dSk ) k = 0 1 2 : : : n . American algorithm. vn(x) = g (x) 1 p vk (x) = max 1 + r (~vk+1 (ux) + q vk+1 (dx)) g(x) ~ Then vk (Sk ) is the value of the option at time k. Thus. the holder of the derivative security can “exercise” and receive payment g (Sk ). 77 . Define by backward recursion: vn (x) = g (x) 1 ~ vk (x) = 1 + r pvk+1 (ux) + qvk+1(dx)]: ~ Then vk (Sk ) is the value of the option at time k. and the wealth process value at that time must equal the value of the derivative security. and the hedging portfolio is given by vk+1 (uSk ) .

Set v2 (x) = g(x) = (5 .1 See Fig. 4)+ 2 5 2 = maxf1:36 1g = 1:36 Let us now construct the hedging portfolio for this option. 2)+ 2 5 2 = maxf2 3g = 3:00 v0 (4) = max 4 1 :(0:4) + 1 :(3:0) (5 . 2 v1 (8) = max 4 1 :0 + 1 :1 (5 . 5. x)+ . Begin with initial wealth X 0 0 as follows: = 1:36. 4 0) 4 3 0 + 1:70 =) 0 = . Example 5.0:43 . S0 Then 1 ~ ~ 2 = 4 u = 2 d = 1 r = 4 p = q = 1 n = 2.3 0 + 1:70 =) 0 = . 0S0 ) 8 0 + 5 (1:36 . Compute 0:40 = = = = 3:00 = = = = v1 (S1 (H)) S1 (H) 0 + (1 + r)(X0 . 8)+ 2 5 2 = max 2 0 5 = 0:40 v1 (2) = max 4 1 :1 + 1 :4 (5 .0:43 v1 (S1 (T)) S1 (T) 0 + (1 + r)(X0 .1: Stock price and final value of an American put option with strike price 5.1. 0S0 ) 2 0 + 5 (1:36 . 4 0) 4 .78 S2 (HH) = 16 v2 (16) = 0 S (H) = 8 1 S2 (HT) = 4 S =4 0 S1 (T) = 2 S2 (TH) = 4 v (4) = 1 2 S2 (TT) = 1 v (1) = 4 2 Figure 5.

1 1 (T) = . The value satisfies Xk g(Sk) k = 0 1 : : : n. f 1 IE( 1 + r vk+1(Sk+1)jF k ] < vk (Sk ) . the value of the European put. 2 1(T)) 1:5 1(T ) + 3:75 =) 1(T) = . (k+1) vk+1 (Sk+1 )jF k ] < (1 + r) k vk (Sk ) . we would have 1 = 1:5 1(T ) + 2:5 =) 4 = . 1(T)S1 (T)) 5 4 1(T ) + 4 (3 .CHAPTER 5. Ck . .2 Value of Portfolio Hedging an American Option Xk+1 = = k Sk+1 + (1 + r)(Xk . k Sk ) (1 + r)Xk + k (Sk+1 . The value process is the smallest process with these properties. (1 + r)Sk ) . When do you consume? If f IE ((1 + r) or. Stopping Times and American Options Using 0 79 = .0:43 results in X1 (H) = v1(S1 (H)) = 0:40 X1 (T ) = v1 (S1 (T)) = 3:00 (Recall that S1 (T) = 2): Now let us compute 1 1 = = = = 4 = = = = We get different answers for 1 v2 (4) S2 (TH) 1(T) + (1 + r)(X1 (T ) . 1 (T )S1 (T )) 5 1(T) + (3 .1:83 v2 (1) S2 (TT ) 1(T ) + (1 + r)(X1 (T) .1:5 1(T) + 2:5 =) 1 (T) = .1 5. The discounted value of the portfolio is a supermartingale. 2 1(T)) 4 .1:5 1(T ) + 3:75 =) 1(T ) = . Ck is the amount “consumed” at time k. (1 + r)Ck Here. equivalently.0:16 (T)! If we had X1 (T ) = 2.

we have v1 (S1(T )) = g (S1(T )). A stopping time is a random variable f! 2 P) be a probability space and let fF k gn=0 be a filtrak : !f0 1 2 : : : ng f1g with the property that: (!) = kg 2 F k 8k = 0 1 : : : n 1: q= ~ 1. Thereafter. then the seller of the option can consume to close the gap. define (!) = minfk Sk (!) = m2 (!)g . he can ensure that X k = vk (Sk ) for all k. he uses the usual hedging portfolio k k vk = vk+1 (uSu ) . so p = ~ 2 4 v0 v1 v2 be the value functions defined for the American put with strike price 5.1. Define Let (!) = minfk vk(Sk ) = (5 . If the owner of the option does not exercise it at time one in the state ! 1 = T . It is optimal for the owner of the American option to exercise whenever its value vk (Sk ) agrees with its intrinsic value g (S k ). then the seller can consume 1 at time 1. Sk )+ g: The stopping time corresponds to “stopping the first time the value of the option agrees with its intrinsic value”.2 Consider the binomial model with n = 2 S 0 = 4 u = 2 d = 1 r = 1 . v1 (S1(T )) = 3 v2(S2(TH )) = 1 and v2(S2 (TT )) = 4.80 and the holder of the American option does not exercise. where vk is the value defined by the American algorithm in Section 5.1 (Stopping Time) Let ( F tion. Therefore.3 (A random time which is not a stopping time) In the same binomial model as in the previous example. 2 Example 5.d k In the example. Definition 5. It is an optimal exercise time. 1 f 1 IE 1 + r v2(S2 )jF 1 ](T ) = 4 2 :1 + 1 :4 2 5 h i = = v1 (S1(T )) = 4 5 5 2 2 3 so there is a gap of size 1. We note that (!) = We verify that is indeed a stopping time: 1 2 if ! if ! 2 AT 2 AH f! (!) = 0g = 2 F 0 f! (!) = 1g = AT 2 F 1 f! (!) = 2g = AH 2 F 2 Example 5. By doing this. In the previous example. )S+1 (dSk ) ( .

4 In the binomial model considered earlier. In other words.e. but the set fTH g is not.1 (Value of Stochastic Process at a Stopping Time) If ( F P) is a probability space. and is k k a stopping time with respect to the same filtration.2 Let that be a stopping time. Indeed. which we denote by F . fF k gn=0 is a filtration under F . random variable is given by 81 4 80 < (!) = : 1 2 stops when the stock price reaches its minimum value. This if ! if ! if ! 2 AH = TH = TT We verify that is not a stopping time: f! (!) = 0g = AH 62 F 0 f! (!) = 1g = fTH g 62 F 1 f! (!) = 2g = fTT g 2 F 2 5. let = minfk vk (Sk ) = (5 .. (!) = The set fHT g is determined by time . We say that a set A is determined by time provided A \ f! (!) = kg 2 F k 8k: The collection of sets determined by is a -algebra.3 Information up to a Stopping Time Definition 5. 1 2 if ! if ! 2 AT 2 AH fHT g \ f! (!) = 0g = 2 F 0 fHT g \ f! (!) = 1g = 2 F 1 fHT g \ f! (!) = 2g = fHT g 2 F 2 but The atoms of F are fT H g \ f! (!) = 1g = fTH g 62 F 1 : fHT g fHH g AT = fT H T T g: Notation 5. then X is an F -measurable random variable whose value at ! is given by X (!) = X (!) (!): 4 . Sk )+ g i. Example 5. fXk gn=0 is a stochastic process adapted to this filtration. Stopping Times and American Options where m2 = min0 j 2 Sj .CHAPTER 5.

5. " 4 2 S F (HH) = 5 2 # 4 2 S F (HT) = 5 2 # 4 5 4 5 2 S2 (HH) S2 (HT ) 2 2 2 AT . Y 0 = IEY0 IEY is a supermartingale. If fYk F k gk=0 is a martingale.82 Theorem 3.5 In the example 5. there is a nonrandom number n such that n If almost surely.17 (Optional Sampling) Suppose that fYk F k g1 (or fYk F k gn=0 ) is a submartink=0 k gale. Under the risk-neutral probability measure. Let and be bounded stopping times. If fYk F k gk=0 1 Example 5. We compute 4 e IE The atoms of F are fHH g " 4 5 2 S2 F : # fHT g e IE e IE and for ! " and AT .2) e IE " 4 5 2 S2 F (!) = 4 5 # ( ) ! S (! ) (!): . implies Y = IE (Y jF ). Therefore. then implies Y IE (Y jF ). e IE " 4 5 2 S2 F (!) = # 4 2 S (T H) + 1 4 2 5 5 2 = 1 2:56 + 1 0:64 2 2 = 1:60 1 2 S2 (T T) In every case we have gotten (see Fig. we define (!) = 2 for all ! 2 . and in particular. then n almost surely..k Sk is a martingale. Y IE (Y jF ): Taking expectations.4 considered earlier. we obtain IEY IEY . then 1 .e. i. the discounted stock price process ( 5 ).

Stopping Times and American Options 83 (16/25) S (HH) = 10.64 Figure 5.24 2 (4/5) S (H) = 6.2: Illustrating the optional sampling theorem.56 (4/5) S (T) = 1.56 S =4 0 (16/25) S2 (TH) = 2.CHAPTER 5. .40 1 (16/25) S2 (HT) = 2.60 1 (16/25) S2 (TT) = 0.

84 .

where the maximum is over all stopping times satisfying k almost surely. 1: !k H ) . Sk+1 (!1 !k T ) k+1 !1 85 .Chapter 6 Properties of American Derivative Securities 6. Vk+1(!1 :: :: :: !k T ) k = 0 1 : : : n . (a) The value Vk of the security at time k is f Vk = max (1 + r)k IE (1 + r) G jF k ] . (c) Any stopping time which satisfies f V0 = IE (1 + r) G ] . he receives the payment Gk .1 An American derivative security is a sequence of non-negative random variables fGk gn=0 such that each Gk is F k -measurable.1 The properties Definition 6. and if he does. The owner of an American derivative security can k exercise at any time k.k Vk gn=0 is the smallest supermartingale which satisfies k Vk Gk 8k almost surely. is an optimal exercise time. (d) The hedging portfolio is given by k (!1 V ( : : : !k ) = Sk+1(!1 :: :: :: !k H ) . (b) The discounted value process f(1 + r). In particular = minfk Vk = Gk g 4 is an optimal exercise time.

k Yk gn=0 is a supermartingale. .s.2 Proofs of the Properties Let fGk gn=0 be a sequence of non-negative random variables such that each Gk is F k -measurable.e.18 Proof: Take Vk Gk for every k. If he does not. h i Lemma 2. Yk Vk k = 0 1 : : : n a. we have f IE (1 + r) (k+1) Vk+1jF k ] f f IE IE (1 + r) G jF k+1]jF k f IE (1 + r) G jF k ] f max IE (1 + r) G jF k ] Tk = (1 + r) k Vk : = = .. 6. we have Vk (!) = Gk (!). Then the owner of the derivative security should exercise it. Define also f Vk = (1 + r)k max IE (1 + r) G jF k ] : T 4 . 2 k Lemma 2. . 2 Tk to be the constant k.k Vk gn=0 is a supermartingale. k Proof: Let fh (1 + r) (k+1) Vk+1 = IE (1 + r) . .20 If fYk gn=0 is another process satisfying k Yk Gk k = 0 1 : : : n and f(1 + r). then k a. 2 .86 (e) Suppose for some k and ! .s. Lemma 2.19 The process f(1 + r). 1 + r IE Vk+1jF k ](!) and still maintain the hedge. G jF k+1 : . attain the maximum in the definition of V k+1 . k Define Tk to be the set of all stopping times satisfying k n almost surely. . then the seller of the security can immediately consume 1 f Vk (!) .. i. i Because is also in T k .

k Yk gn=0 implies k 87 f IE (1 + r) Y jF k ] (1 + r) k Yk 8 2 Tk : .21 Define f (1 + r)k max IE (1 + r) Y jF k ] Tk (1 + r) k (1 + r)k Yk . IE (1 + r) . (k+1) Vk+1jF k ] o : Since f(1 + r).e. for each fixed (!1 : : : !k) we have Xk (!1 : : : !k ) = Vk (!1 : : : !k ): Proof: We need to show that Xk+1 (!1 : : : !k H ) = Vk+1(!1 : : : !k H ) Xk+1(!1 : : : !k T ) = Vk+1(!1 : : : !k T ): We prove the first equality. 2 . 2 k 1 f Ck = Vk .k Vk gn=0 is a supermartingale. Ck (!1 : : : !k ) f = 1 IE Vk+1 jF k ](!1 : : : !k ) 1+r 1 p = 1 + r (~Vk+1 (!1 : : : !k H ) + q Vk+1 (!1 : : : !k T )) : ~ . . Sk+1 (!1 !k T ) k+1 (!1 Then Xk = Vk 8k: for some We proceed by induction on k. Properties of American Derivative Securities Proof: The optional sampling theorem for the supermartingale f(1 + r) . = Yk : Lemma 2. Therefore. i. Note first that Vk (!1 : : : !k ) .CHAPTER 6. 1 + r IE Vk+1 jF k ] n f = (1 + r)k (1 + r) k Vk . The induction hypothesis is that X k = Vk k 2 f0 1 : : : n . 1g. C k must be non-negative almost surely. . Vk+1 (!1 :: :: :: !k T ) : !k H ) . Define k k (!1 Set X0 = V0 and define recursively Xk+1 = k Sk+1 + (1 + r)(Xk . k Sk ): V : : : !k ) = Sk+1(!1 :: :: :: !k H ) .. f Vk = (1 + r)k max IE (1 + r) G jF k ] T . the proof of the second is similar. Ck .

start with wealth V 0 = n=0 V0 . Vk+1(T ) (S (H ) . Thus a compound European derivative security is specified by the process fCj gn=0 . Vk(+1 (!1 : : : !k T ) : : : !k ) = (j) k = 0 : : : j . A compound European derivative security consists of n + 1 different simple European derivative securities (with the same underlying stock) expiring at times 0 1 : : : n.3 Compound European Derivative Securities In order to derive the optimal stopping time for an American derivative security. In P (j) other words. Here is how we can hedge a short position in the j ’th European derivative security. we will suppress these symbols. and using the portfolio ( 0 time j we have wealth Cj . Ck ) Vk+1 (H ) . At each time k 2 f0 1 : : : n . 1: () Sk+1 (!1 : : : !k H ) . and the hedging portfolio for that security is given by (j ) (! k 1 j) j) Vk(+1 (!1 : : : !k H ) . Vk+1 (T ))~ + pVk+1(H ) + q Vk+1 (T ) q ~ ~ Vk+1 (H ): 6. first make the j payment Ck and then use the portfolio k = k (k+1) + k (k+2) + : : : + k (n) .. Superpose the hedges for the individual payments. 1g. k Hedging a short position (one payment). For example. the process fCj gn=0 is adapted to the filtration j j fF k gn=0. d)Sk +~Vk+1 (H ) + q Vk+1 (T ) p ~ (Vk+1 (H ) . it will be useful to study compound European derivative securities. the security that expires at time j has payoff Cj . Ck = 1 + r (~Vk+1 (H ) + qVk+1 (T )) : ~ We compute Xk+1 (H ) = = = = = k Sk+1 (H ) + (1 + r)(Xk . where each Cj is F j -measurable. the last equation can be written simply as 1 p Vk .88 Since (!1 : : : !k ) will be fixed for the rest of the proof. we can ensure that at j 1 .e. (1 + r)S ) k k (u . Skj+1 (!1 : : : !k T ) (j ) (j ) Thus. i. (1 + r)S ) k Sk+1 (H ) . Ck . Sk+1 (T ) k+1 +(1 + r)(Vk . Hedging a short position (all payments). k Sk ) Vk+1 (H ) . starting with wealth V 0 . which are also interesting in their own right. ::: (j ) ). The value of European derivative security j at time k is given by f Vk(j) = (1 + r)k IE (1 + r) j Cj jF k ] k = 0 : : : j . Vk+1(T ) (uS .

CHAPTER 6. Properties of American Derivative Securities

89

corresponding to all future payments. At the final time n, after making the final payment Cn , we will have exactly zero wealth. Suppose you own a compound European derivative securityfC j gn=0 . Compute j

V0 =

n X j =0

2n 3 X f V0(j ) = IE 4 (1 + r) j Cj 5
;

j =0

;1 and the hedging portfolio is f k gn=0 . You can borrow V0 and consume it immediately. This leaves k you with wealth X 0 = ;V0 . In each period k, receive the payment Ck and then use the portfolio ; k . At the final time n, after receiving the last payment Cn, your wealth will reach zero, i.e., you will no longer have a debt.

6.4 Optimal Exercise of American Derivative Security
In this section we derive the optimal exercise time for the owner of an American derivative security. Let fGk gn=0 be an American derivative security. Let be the stopping time the owner plans to k use. (We assume that each Gk is non-negative, so we may assume without loss of generality that the owner stops at expiration – time n– if not before). Using the stopping time , in period j the owner will receive the payment

Cj = I

f

=j

g

Gj :

In other words, once he chooses a stopping time, the owner has effectively converted the American derivative security into a compound European derivative security, whose value is

V0( )

2n 3 X f = IE 4 (1 + r) j Cj 5 2j=0 3 n X f = IE 4 (1 + r) j I =j Gj 5
; ;

f = IE (1 + r) G ]:
;

j =0

f

g

The owner of the American derivative security can borrow this amount of money immediately, if he chooses, and invest in the market so as to exaclty pay off his debt as the payments fC j gn=0 are j received. Thus, his optimal behavior is to use a stopping time Lemma 4.22 which maximizes V 0 .

( )

V0( ) is maximized by the stopping time
= minfk Vk = Gk g:

Proof: Recall the definition

f V0 = max IE (1 + r) G ] = max V0( T T
4 ; 2

)

0

2

0

90

f Let 0 be a stopping time which maximizes V0 , i.e., V0 = I (1 + r); G 0 : Because f(1 + r);k Vk gn=0 E k is a supermartingale, we have from the optional sampling theorem and the inequality V k Gk , the following:
( )
0

h

i

V0

f IE (1 + r) 0 V 0 jF 0 i fh = IE (1 + r) 0 V 0 i fh IE (1 + r) 0 G 0 = V0:
; ; ; ;

h

i

Therefore, and

f f V0 = IE (1 + r) 0 V 0 = IE (1 + r) 0 G 0
;

h

i

h

i

V 0=G
0

0 a.s.

We have just shown that if

attains the maximum in the formula

f V0 = max IE (1 + r) G ] T
; 2

0

(4.1)

then But we have defined and so we must have
0

V 0=G

0 a.s.

n almost surely. The optional sampling theorem implies
;

= minfk Vk = Gk g
;

(1 + r)

G

= (1 + r) V i fh IE (1 + r) 0 V 0 jF i fh = IE (1 + r) 0 G 0 jF :
; ;

Taking expectations on both sides, we obtain

f IE (1 + r)

h

;

G

i fh i IE (1 + r) 0 G 0 = V0 :
;

It follows that also attains the maximum in (4.1), and is therefore an optimal exercise time for the American derivative security.

Chapter 7

Jensen’s Inequality
7.1 Jensen’s Inequality for Conditional Expectations
Lemma 1.23 If ' : IR!IR is convex and IE j'(X )j < 1, then

IE '(X )jG] '(IE X jG]):
For instance, if G

=f

g '(x) = x2:
IEX 2 (IEX )2:

Proof: Since ' is convex we can express it as follows (See Fig. 7.1):

'(x) = max h(x): h '
h is linear
Now let h(x) = ax + b lie below '. Then,

IE '(X )jG]

IE aX + bjG] = aIE X jG] + b = h(IE X jG])
max h(IE X jG]) h '

This implies

IE '(X )jG]

= '(IE X jG]):

h is linear

91

92

ϕ

Figure 7.1: Expressing a convex function as a max over linear functions. Theorem 1.24 If fYk gn=0 is a martingale and k Proof: is convex then f'(Y k )gn=0 is a submartingale. k

IE '(Yk+1)jF k ]

'(IE Yk+1jF k ]) = '(Yk ):

7.2 Optimal Exercise of an American Call
This follows from Jensen’s inequality. Corollary 2.25 Given a convex function g : 0 1)!IR where g (0) = 0. For instance, g (x) = (x ; K )+ is the payoff function for an American call. Assume that r 0. Consider the American derivative security with payoff g (S k ) in period k. The value of this security is the same as the value of the simple European derivative security with final payoff g (S n), i.e., f f IE (1 + r) n g (Sn)] = max IE (1 + r) g (S )]
; ;

where the LHS is the European value and the RHS is the American value. In particular optimal exercise time. Proof: Because g is convex, for all

= n is an

2 0 1] we have (see Fig. 7.2): g( x) = g ( x + (1 ; ):0) g (x) + (1 ; ):g (0)
=

g (x):

CHAPTER 7. Jensen’s Inequality

93

( λ x, λ g(x))

(x,g(x))

x ( λ x, g( λ x))
Figure 7.2: Proof of Cor. 2.25 Therefore,

1 g 1 + r Sk+1

1 g (S ) 1 + r k+1
;

and

optional sampling theorem implies

1 g (S )jF 1 + r k+1 k 1 f (1 + r) k IE g 1 + r Sk+1 jF k f 1 (1 + r) k g IE 1 + r Sk+1 jF k = (1 + r) k g (Sk ) So f(1 + r) k g (Sk )gn=0 is a submartingale. Let be a stopping time satisfying 0 k

f IE (1 + r)

h

;

(k+1) g (Sk+1)jF k

i

f = (1 + r) k IE
; ;

;

;

n. The

f (1 + r) g (S ) IE (1 + r) n g (Sn)jF ] :
; ; ;

Taking expectations, we obtain

f IE (1 + r) g(S )]

f f IE IE (1 + r) n g(Sn)jF ] f = IE (1 + r) n g (Sn)] :
; ; ;

Therefore, the value of the American derivative security is
;

f f max IE (1 + r) g (S )] IE (1 + r) n g (Sn )] f f max IE (1 + r) g (S )] = IE (1 + r) n g (Sn )] :
; ;

and this last expression is the value of the European derivative security. Of course, the LHS cannot be strictly less than the RHS above, since stopping at time n is always allowed, and we conclude that

94
S2 (HH) = 16

S (H) = 8 1 S2 (HT) = 4 S =4 0 S1 (T) = 2 S2 (TH) = 4

S2 (TT) = 1

Figure 7.3: A three period binomial model.

7.3 Stopped Martingales
Let fYk gn=0 be a stochastic process and let k stopped process
^

be a stopping time. We denote by

fY k gn=0 the k
^

Yk (!) (!) k = 0 1 : : : n:
(!) = 1 2
if if

Example 7.1 (Stopped Process) Figure 7.3 shows our familiar 3-period binomial example. Define

Then

8 S (HH) = 16 > 2 < ) S2^ (! ) (!) = > S2 (HT= = 4 S1 (T ) 2 : S1(T ) = 2

!1 = T !1 = H:
if if if if

! = HH ! = HT ! = TH ! = T T:

Theorem 3.26 A stopped martingale (or submartingale, or supermartingale) is still a martingale (or submartingale, or supermartingale respectively). Proof: Let fYk gn=0 be a martingale, and be a stopping time. Choose some k 2 f0 1 k The set f kg is in F k , so the set f k + 1g = f kgc is also in F k . We compute

: : : ng.

IE Y(k+1) jF k = IE I k Y + I k+1 Yk+1 jF k = I k Y + I k+1 IE Yk+1 jF k ] = I k Y + I k+1 Yk = Yk :
^ f g f g f f g g f f g g ^

h

i

h

i

CHAPTER 7. Jensen’s Inequality

95

96 .

1) Its analogue in continuous time is k=0 Brownian motion. . and on each toss.. Define 2 Yj (!) = Mk = ( .g.1 First Passage Time Toss a coin infinitely many times. ! = (TTTT : : : )). as is the probability of T . Define j =1 Yj k = 1 2 : : : = minfk 0 Mk = 1g: If Mk never gets to 1 (e. It is shown in a Homework Problem that fM k g1 and fNk g1 where k=0 k=0 Nk = exp Mk .2 is almost surely finite is called the first passage time to 1. It is the first time the number of heads exceeds by one the number of tails. Assume the tosses are independent. The random variable 8. 8. !) k . then = 1. k log e + e 2 = e Mk 2 e +e 97 ( .Chapter 8 Random Walks 8. Then the sample space is the set of all infinite sequences ! = (!1 !2 : : : ) of H and T . the probability of H is 1 .1 k X 1 if if !j = H !j = T M0 = 0 The process fMk g1 is a symmetric random walk (see Fig.

2 for an illustration of the various functions involved). (2. so 0 e Mk ^ =1 e . (! ) = 1. we have =.98 Mk k Figure 8. so e +e 2 e +e k . We want to let k!1 in (2.2: Illustrating two functions of 2 eθ e−θ + 1 θ are martingales.1). ^ 2 . Mk^ 1. 8. (Take Mk = .Sk in part (i) of the Homework Problem and take (v).1: The random walk process Mk eθ e−θ + 2 1 θ Figure 8. <1 Furthermore.1) for every fixed 2 IR (See Fig. because we stop this martingale when it reaches 1. < 1: if if ! ( 0 2 e +e. > 0. We consider fixed As k!1.) Since N0 = 1 and a stopped martingale is a martingale. in part 1 = IENk = IE e Mk ^ " ^ 2 e +e k ^ # . but we have to worry a bit that for some sequences ! 2 .

.1): IE e Mk^ " Letting k!1. We k=0 have just shown that these paths collectively have no probability. ^ 2 e +e . <1 = 1: Recall Equation (2. (2. . we obtain 2 e +e k ^ # . (In our infinite sample space . Random Walks and 99 0 e Mk lim e Mk k! 1 ^ 2 e +e k ^ k ^ . we have so we can let #0 in (2.2) 2 (0 1]. =e : . using the Bounded Convergence Theorem again.2). . We know there are paths of the symmetric random walk fMk g1 which never reach level 1. 2e + = 0 . and we rewrite that equation as IE e +e 2 . I f < 1g (2. to conclude h i IE If < 1g = 1 IP f < 1g = 1: 2 0 e e +e . : Solution: e + e .2=0 (e )2 .3) 8. We therefore do not need the indicator I f < 1g in (2. I f < 1g e i. We want to find > 0 so that 2 = e +e . e: if if In addition. = ( 0 e 2 e +e.CHAPTER 8..3 The moment generating function for Let 2 (0 1) be given.2). =1 = 1: 2 IE e e + e For all . and using the Bounded Convergence Theorem. each path individually has zero probability).e.

1. 2 p 1. 2 : 2 We want > 0. 1) p 2 ! p 2 2 : . )2 < (1 . =e . p 2 : > 0: 2 Recall Equation (2. p this becomes IE = 1. 1.3): IE With 2 (0 1) and > 0 related by e e +e . . p 1. 2 = e +e . 1. = 1. 2 < p 1. 2 1.4 Expectation of Recall that IE so = 1. 1. p 2 0 < < 1: (3. 1 . so we must have e < 1. 1. = dd 1 . d IE d = 1 . 1. 2 < 1 We take the negative square root: p e = 1. = IE ( p 2 0< <1 .1) We have computed the moment generating function for the first passage time to 1. 0 < (1 . 2 . ) < 1 . so . 1. 8.100 e =1 . Now 0 < < 1.p 1 . < 1.

3). m = minfk 4 0 Mk = mg m = 1 2 : : : IE 2. 1 = 1. 2 1. i. 1) = 1 . 1 is the same as the distribution of 1 (see Fig.6 General First Passage Times Define Then 2 .e. 1 is the number of periods between the first arrival at level 1 and the first arrival at level 2. The distribution of 2 . i. 1. p 2 2 (0 1): . 8. we can let 101 IE ( to obtain .p 1 .. Random Walks Using the Monotone Convergence Theorem. p "1 in the equation 2 2 IE = 1: = minfk Mk = 1g 4 Thus in summary: IP f < 1g = 1 IE = 1: 8.e. 8. this Markov property implies the Strong Markov property: IE random variable depending only on = IE same random variable M +1 M +2 : : : j F ] j M ]: for any almost surely finite stopping time .5 The Strong Markov Property The random walk process fMk g1 is a Markov process..CHAPTER 8. k=0 IE random variable depending only on = IE same random variable Mk+1 Mk+2 : : : j F k ] jMk ] : In discrete time.

2 : IE 2 = IE 1 : 1 . 1. IE 2 jF 1 ] = IE = 1 1 2. IE m = 1. 1 jM 1 ] IE 2 1 ] (M 1 = 1 not random ) ! p 1 . q = 1 . and payoff function (5 . = 1. 1 IE IE jF 1 jF 1 ] = = = Take expectations of both sides to get (taking out what is known) 1 2.7 Example: Perpetual American Put 1 Consider the binomial model. and thus ~ neutral probabilities are p = 2 ~ 2 r= 1 . For τ2 2 (0 1). with u = 2 d = 2 1 . 1 2. 1 . p 2 ! p ! 2 2 In general. 1. (strong Markov property) 1 1.3: General first passage times. 1.102 Mk k τ1 τ 2 − τ1 Figure 8. The risk k 4 Sk = S0 uMk . S )+ . p ! 2 m 2 (0 1): 8.

2)+ IE ( 5 ) . and stop the first time the price falls to 2.. if S0 = 2j for some j 1.1 )+ h4 i = (5 . S . then j . = 0 V ( 0 ) = 1. 2)+ IE ( 5 ) . f 4 (5 .. 1 j = 1 2 3 ::: .1) = 3:( 1 )j 1 : 2 We define v(2j ) = 3:( 1 )j 2 4 .1 .2 ) = (5 . i. This requires 2 down steps.2 = 3:( 1 )2 = 3 : 2 4 h i .(j. and the value of the put is 3 if S0 = 4. 1)+IE ( 4 ) .1) = IE (1 + r) .. 0 where f Mk is a symmetric random walk under the risk-neutral measure.CHAPTER 8.2 1 = 4:( 2 )2 = 1: h i Suppose instead we start with S 0 = 8.1g: 4 Rule 2: Stop as soon as stock price falls to 1. = 3: 3 = 2: Value of Rule 2: 1 . 1 . 2 In general. i. 2 = minfk Mk = . at time . 1 Rule 1: Stop as soon as stock price falls to 2. and we stop when the stock price falls to 2.e. denoted by IP . Value of Rule 1: f V ( . . ( 4 )2 5 4 5 q f 5 V ( . so the value of this rule with this initial stock price is This suggests that the optimal rule is Rule 1.e. 1 down steps will be required and the value of the option is fh 4 i (5 . i. = minfk Mk = .2g: 4 f f Because the random walk is symmetric under I . stop (exercise the put) as soon as the stock price falls to 2. Random Walks 103 Suppose S0 = 4. 2)+IE ( 5 ) .1 (5 .e. at time . Here are some possible exercise rules: Rule 0: Stop immediately.m has the same distribution under I as the P P stopping time m in the previous section. This observation leads to the following computations of value.

2 : : : 4 Proposed exercise rule: Exercise the put whenever the stock price is at or below 2. the initial time is no different from any other time. 5 k=0 (c) fv (Sk )gk=0 is the smallest process with properties (a) and (b). How do we recognize the value of an American derivative security when we see it? There are three parts to the proof of the conjecture. with a possibly different j . We must show that v(Sk ) By assumption.1 . then v (2j ) = 3:( 1 )j 1 and 2 2 v (2j +1 ) + 2 v (2j 1) 5 5 1 1 = 2 :3:( 2 )j + 2 :3:( 2 )j 2 5 5 2 = 3: 5 : 1 + 2 ( 1 )j 2 4 5 2 = 3: 1 :( 1 )j 2 2 2 = v (2j ): . . We must show that 2 v(2j ) 5 v (2j +1) + 2 v (2j 1): 5 If j 2. . In this case. 2j j = 1 0 . we exercise If S0 = 2j for some j immediately. . 2j (5 . Just check that = 2j . and the value of the put is v (2j ) = 5 . Proof: (b). .104 1. 1 (5 . f 5 IE 4 v (Sk+1)jF k = 4 : 1 v (2Sk ) + 4 : 1 v ( 1 Sk ): 5 2 5 2 2 h i . we shall only consider initial stock prices of the form S 0 always of the form 2j . We must show: (a) v (Sk ) (5 . . 2j )+ This is straightforward. so Sk is v(2j ) = 3:( 1 )j 2 4 4 . The value of this rule is given by v (2 j ) as we just defined it. 2j )+ for for j 1 j 1: v(2j ) = 5 . Since the put is perpetual. Sk )+ 8k o n (b) ( 4 )k v (Sk ) is a supermartingale. 1 1 Note: To simplify matters. Proof: (a). then the initial price is at or below 2. This leads us to make the following: Conjecture 1 The value of the perpetual put at time k is v (S k ). Sk = 2j for some j .

5 1 We must show that Yk v (Sk ) 8k: Y0 v(S0) For j (7. 5 0. 2j 1 ) 5 5 2 = 4 .2) can be modified to prove (7. There is a gap of size 1. This concludes the proof of (b). since the put is perpetual. 2j = (5 . so it will suffice to show (7.CHAPTER 8.e. = 3=5 + 8=5 1 = 2 5 < v (2) = 3 If j There is a gap of size 4 .. S0 4. i. Random Walks If j 105 = 1. 1. 2j )+ Y0 (5 . 2j < v (2j ) = 5 . . Suppose fYk gn=0 is some other process satisfying: k (a’) (b’) Yk (5 . 2j +1 ) + 2 (5 . 2j and 2 v (2j +1 ) + 2 v (2j 1) 5 5 = 2 (5 . then v (2j ) = v (2) = 3 and 2 v (2j +1 ) + 2 v (2j 1) 5 5 2 = 5 v (4) + 2 v (1) 5 2 :3: 1 + 2 :4 = 5 2 5 . Let = minfk Sk = 2g = minfk Mk = j . 2j )+ = v (S0): Suppose now that S 0 = 2j for some j 2. then v (2j ) = 5 . Sk )+ 8k f( 4 )k Yk gk=0 is a supermartingale.1).1) Actually. 1g: . 2j : .2) be any number of the form 2j . so if S0 = 2j for some j 1. then (a’) implies v (2j ) = 5 . 5 (4 + 1)2j 1 = 4 . every time k is like every other time.2) provided we let S0 in (7. Proof: (c). the proof we give of (7. With appropriate (but messy) conditioning on F k . .

not just at points of the form 2 j for integers j . h i . 4 Because f( 5 )k Yk g1 is a supermartingale k=0 4 Y0 IE ( 5 ) Y h i IE ( 4 ) (5 .106 Then v(S0) = v (2j ) = 3:( 1 )j 1 h 4 2 +i = IE ( 5 ) (5 . 8. In this case. This function should satisfy the conditions: v (x) (K . x 0 < x 3: 6 x v (x) (5 . x)+ 8x v (x) 4 1 1 x 5 2 v (2x) + 2 v ( 2 ) for every x except for 2 < x < 4. In such a situation.8 Difference Equation If we imagine stock prices which can fall at any point in (0 1).4): (a) (b) ( x 3 5 . 1 ~ ~ (b) v (x) 1+r pv (ux) + q v (dx)] 8x (c) At each x. S )+ = v(S0): 5 h i Comment on the proof of (c): If the candidate value process is the actual value of a particular exercise rule. either (a) or (b) holds with equality. or later) that the stock price is 2 or less. This suggests the formula v (x) = We then have (see Fig. (a) In the example we worked out. defined for all x > 0. S ) : . we need only verify properties (a) and (b). we constructed v so that v (S k ) is the value of the put at time k if the stock price at time k is S k and if we exercise the put the first time (k. 8. x)+ 8x. 2j : . we have For 6 j 1 : v(2j ) = 3:( 1 )j 1 = 2j 2 For j 1 : v (2j ) = 5 . then (c) will be automatically satisfied. which gives the value of the perpetual American put when the stock price is x. then we can imagine the function v (x).

p 2 0 < < 1: We will use this moment generating function to obtain the distribution of . in which an analogue of (a) or (b) holds with equality at every point. We first obtain the Taylor series expasion of IE as follows: .4: Graph of v (x). then (a) holds with equality.9 Distribution of First Passage Times Let fMk g1 be a symetric random walk under a probability measure IP . Random Walks 107 v(x) 5 (3. then both (a) and (b) are strict. This is an artifact of the discreteness of the binomial model.CHAPTER 8. 1. 8.2) 5 x Figure 8. Defining = minfk 0 Mk = 1g we recall that IE = 1. then (b) holds with equality: =4 5 6 = x: If 3 < x < 4 or 4 < x < 6. 4 1 v (2x) + 1 v ( x ) 5 2 2 2 1 12 1 6 2 2x + 2 x 6. This artifact will disappear in the continuous model. with M0 k=0 = 0. Check of condition (c): If 0 < x If x 3.

2 : ! j But also. x) 5 f (0) = 3 2 f (x) = 8 8 ::: : f (j)(x) = 1 3 : :2j (2j . 2j The Taylor series expansion of f (x) is given by f (x) = 1 . 1)! x .108 f (x) = 1 . 1) 1 2j . 1 . 00 000 . j =2 1 2j 1 1 2 (j . 1)! . 1) 2j . x) 2 f (0) = 1 4 4 3 (1 . 1. 000 p . 1 . 0 00 . 1g: . x) 2 f (0) = 2 3 f (x) = 1 (1 . : = 1 3 : :2j (2j . (2j . x) f (j)(0) = 1 3 : : : (2j . j =2 1 2 2j 1 IP f . 2) j 1)! 2j 1 (2j . 2)! j j !(j . = 1 f ( 2) = 2 + p 2 X 1 2j 1 . x X 1 (j) j = j ! f (0)x 1 p j =0 = X 1 X = x+ 2 1 j =1 1 2j 1 2 . . x f (0) = 0 1 1 1 f (x) = 2 (1 .(2j . 3) (1 . 3) 0 . 2 xj : j ! So we have IE = 1. (2j 1) 2 . 2)! 1 = 2 (j . IE = X j =1 = 2j . (j . 3) : 2 42j :1:(: .

8.) = 1 and .1 3. 1) j .CHAPTER 8. 2 j = 2 3 : : : 2 (j .1 double count all those for which M 2j .5. (See Figures 8.6.5: Reflection principle. 1g = 1 2 ! 1 2j 1 1 2j . = 2. IP f = 1g = IP f = 2j . 8. 1.6: Example with j Therefore.10 The Reflection Principle To count how many paths reach level 1 by time 2j . Figure 8. Random Walks 109 Figure 8. count all those for which M 2j .

.1g] IP fM2j 3 = . 2)] 2 j !(j . . 1)! 1 2j 3 2 (j . 2)! 2 j !(j . . IP f 2j .110 In other words. .1g . 2)] 1 2j 1 (2j . (2j . 3)! 1 2 j !(j . IP f = 2j .1g (2j . 1 . IP fM2j 1 = . IP f For j 2j .1g: . . 1)! 4j (j . 1) j . . 3)! . 1)! 2j 1 (2j . . . 3g 1 . 1)!(j . . 1 2j 1 (2j . (2j . . 2) . IP fM2j 1 = . . 1g = IP f = = = = = = = 2j . 1)(2j . IP fM2j 1 = . . 2 : 1 2j 1 1 2 (j . 1 . . . IP fM2j 3 = . 1g = IP fM2j 1 = 1g + 2IP fM2j 1 3g = IP fM2j 1 = 1g + IP fM2j 1 3g + IP fM2j = 1 . . 3)! 2j (2j .3g 2. 1g . 1)! ! 2j . 1) . 1)(2j .1g] . 1)! 2j 1 (2j . 2)! 1 2 j !(j .

3) Remark 9.t.3) are the same. and I is absolutely continuous with P P P f. I and IP are equivalent if and only if f f respect to I P P P f IP (A) = 0 exactly when IP (A) = 0 8A 2 F : 1 f f If I and I are equivalent and Z is the Radon-Nikodym derivative of I w.1 Equation (1.2) and (1. P P f (1.2) IEX = IE XZ ] 8X f 1 IEY = IE Y: Z ] 8Y: (Let X and Y be related by the equation Y = XZ to see that (1. Under this assumption.t. P f Assume that for every A 2 F satisfying IP (A) = 0. f IEX = IE XZ ] for every random variable X for which IE jXZ j < 1.e. f Remark 9. we say that I and IP are equivalent. there is a nonegative random P IP variable Z such that f IP (A) = Z f P P and Z is called the Radon-Nikodym derivative of I with respect to I .2 If I is absolutely continuous with respect to I . 9.1 Radon-Nikodym Theorem f P Theorem 1. we also have I (A) = 0.r..27 (Radon-Nikodym) Let I and I be two probability measures on a space ( F ).1) implies the apparently stronger condition A ZdIP 8A 2 F (1.r.) 111 (1. then Z is the P P P P f Radon-Nikodym derivative of I w.Chapter 9 Pricing in terms of Market Probabilities: The Radon-Nikodym Theorem. Then we say that P f is absolutely continuous with respect to I . i. I . I .1) .

and let IP correspond to probability 1 for 3 3 2 e H and 1 for T . Assume f f P P so that I and I are equivalent. The Radon-Nikodym derivative of I with respect to I is P P f IP (!) Z (!) = IP (! ) : Define the I -martingale P f IP (!) > 0 IP (!) > 0 8! 2 We can check that Zk is indeed a martingale: Zk = IE Z jF k ] k = 0 1 : : : n: 4 f Lemma 2. Then Z(!) = IP (!) . or equivalently. so 2 I ( ) P e ! 9 9 Z(HH) = 4 Z(HT) = 9 Z(TH) = 9 Z(T T) = 16 : 8 8 9.28 If X is F k -measurable.1 (Radon-Nikodym Theorem) Let = fHH HT TH T T g.28 implies that if X is F k -measurable. then for any A 2 F k . f IEX = IE XZ ] = IE IE XZ jF k ]] = IE X:IE Z jF k ]] = IE XZk ]: f IE IA X ] = IE ZkIA X ] A Note that Lemma 2. the set of coin toss sequences of length 2.112 Example 9. Let P correspond to probability 1 for H and 2 for T . Let I be the market probability measure and let P f IP be the risk-neutral probability measure. then I EX Proof: IE Zk+1jF k ] = IE IE Z jF k+1 ]jF k ] = IE Z jF k ] = Zk : = IE XZk ]. Z f XdIP = Z A XZk dIP: .2 Radon-Nikodym Martingales Let be the set of all sequences of n coin tosses.

28) f XdIP Example 9. The probabilities shown f are for I .29 If X is F k -measurable and 0 j k. since Z0 = IEZ = Z e ZdIP = IP ( ) = 1: 9. Pricing in terms of Market Probabilities Z2 (HH) = 9/4 1/3 Z (H) = 3/2 1 1/3 Z =1 0 2/3 Z (T) = 3/4 1 2/3 Z2 (TT) = 9/16 1/3 2/3 Z2 (HT) = 9/8 Z2 (TH) = 9/8 113 Figure 9. So for any A 2 F j . we have (Lemma 2.CHAPTER 9.28) ZA A XZk dIP (Partial averaging) (Lemma 2. P P Lemma 2.1 the values of the martingale Zk . continued) We show in Fig.1: Showing the Zk values in the 2-period binomial model example. .3 The State Price Density Process In order to express the value of a derivative security in terms of the market probabilities. We always have Z0 = 1.2 (Radon-Nikodym Theorem. then j 1 f IE X jF j ] = Z IE XZk jF j ]: 1 Proof: Note first that Z Z Z 1 f IE XZk jF j ]dIP = IE XZkjF j ]dIP A A Zj Z = = j IE XZk jF j ] is F j -measurable. not I . 9. it will be useful to introduce the following state price density process: k = (1 + r) k Zk k = 0 : : : n: .

3 f j Vj gk=0 is a martingale under I . 1 = (1 + r)j sup Z IE (1 + r) Z G jF j ] 2 2 Tj . Tj Remark 9. .28) More generally for 0 j k. 2 2 T0 More generally for 0 f Vj = (1 + r)j sup IE (1 + r) G jF j ] = 1 sup IE G jF j ]: j 2 j n. h i (Lemma 2. Tj j . f V0 = IE (1 + r) k Ck h i = IE (1 + r) k Zk Ck = IE k Ck ]: . . h i (Lemma 2. P j (b) j Vj j Gj 8j (a) .114 We then have the following pricing formulas: For a Simple European derivative security with payoff Ck at time k. as we can check below: P j IE j+1Vj +1 jF j ] = IE IE k Ck jF j +1]jF j ] = IE k Ck jF j ] = j Vj : Now for an American derivative security fGk gn=0 : k 2 f V0 = sup IE (1 + r) G ] T0 = sup IE (1 + r) Z G ] T0 = sup IE G ]: . .29) j Remark 9. f Vj = (1 + r)j IE (1 + r) k Ck jF j j h i = (1 + r) IE (1 + r) k Zk Ck jF j Zj 1 IE C jF ] = k k j .4 Note that f j Vj gn=0 is a supermartingale under I .

expiration time 2.3 (Radon-NikodymTheorem.3 shows the values of + k (5 . P P (c) f j Vj gn=0 is the smallest process having properties (a) and (b). For a European Call with strike price 5. Recall that p = 3 . Pricing in terms of Market Probabilities ζ2(ΗΗ) = 1.2: Showing the state price values k . continued) We illustrate the use of the valuation formulas for 1 European and American derivative securities in terms of market probabilities. j We interpret k by observing that k (! )IP (! ) is the value at time zero of a contract which pays $1 at time k if ! occurs.72 S2 (HT) = 4 S2 (TH) = 4 ζ (ΤΗ) = 0. Sk ) .2. We compute the value of the put under various stopping times : (0) Stop immediately: value is 1. not I .CHAPTER 9. The probabilities shown are for I .20 1 S (H) = 8 1 1/3 S =4 0 2/3 ζ0 = 1.00 2/3 ζ2(ΗΤ) = 0. the value is 1 2 0:72 + 2 1:80 = 1:36: 3 3 3 . The 3 state price values k are shown in Fig. 9. Example 9. we have V2 (HH) = 11 2(HH)V2 (HH) = 1:44 11 = 15:84: V2 (HT) = V2(TH) = V2 (TT ) = 0: V0 = 1 1 15:84 = 1:76: 3 3 2 (HH) V2(HH) = 1:44 11 = 1:20 11 = 13:20 1:20 1 (HH) V1 (H) = 1 13:20 = 4:40 3 Compare with the risk-neutral pricing formulas: V1(H) = 2 V1 (HH) + 2 V1(HT) = 2 11 = 4:40 5 5 5 2 V1 (T ) = 5 V1 (TH) + 2 V1 (T T ) = 0 5 2 V (H) + 2 V (T) = 2 V0 = 5 1 4:40 = 1:76: 5 1 5 Now consider an American put with strike price 5 and expiration time 2. Fig. (1) If (HH) = (HT) = 2 (T H) = (TT ) = 1.72 2 115 1/3 S (T) = 2 1 ζ (Τ) = 0. 9.36 2 S2 (TT) = 1 f Figure 9.6 1 2/3 ζ (ΤΤ) = 0. q = 2 .44 S2 (HH) = 16 1/3 ζ (Η) = 1.

S (T))+ = 1.S (TT)) = 4 2 + ζ (TT) (5 . the value is 1 3 2 3 2 0:72 + 3 1 0:72 + 2 3 3 2 3 1:44 = 0:96 We see that (1) is optimal stopping rule.4 Stochastic Volatility Binomial Model Let be the set of sequences of n tosses. f IP f!1 = H g = p0 ~ f IP f!1 = T g = q0 ~ . not I .72 2/3 + (5 . The probabilities shown are for f I .S (HH)) = 0 2 + ζ (HH) (5 . where for each k. qk = uk u (1 + rk) : ~ .3: Showing the values k (5 . Then I and P f IP are . P P (2) If we stop at time 2. and let 0 < d k are F k -measurable. dk uk rk k k f Let I be the risk-neutral probability measure: P k pk = 1 + r..S2(TH)) = 0. Also let < 1+ rk < uk .S (T)) = 3 1 ζ (T) (5 .44 2 2 Figure 9. dk ~ u d k k . Define f IP !k+1 = H jF k ] = pk ~ f IP !k+1 = T jF k ] = qk : ~ Let I be the market probability measure.S (H)) = 0 1 + ζ (H) (5 .116 + (5 . and assume IP f! g > 0 8! 2 P f IP (!) Z (!) = IP (! ) 8! 2 k n.72 2 2 (5 . 9.S (HH)) = 0 2 2 + (5 .S1(H)) = 0 1 1/3 2/3 + (5 .80 1 1 1/3 + (5-S 0) =1 + ζ0 (5-S 0) =1 2/3 1/3 + (5 .S (HT))+= 0.S (TH))+= 1 2 + ζ2 (TH) (5 .S (TT)) = 1. Sk )+ for an American put.S (HT)) = 1 2 ζ (HT) (5 .d and for 2 equivalent.

1 k = 1 : : : n: We then define the state price process to be k 1 = M Zk k = 0 : : : n: . Pricing in terms of Market Probabilities 117 Zk = IE Z jF k ] k = 0 1 : : : n: We define the money market price process as follows: M0 = 1 Note that Mk is Fk. and k n 1 k gk=0 . Mk = (1 + rk 1)Mk . The self-financing value process (wealth process) Xk+1 = k Sk+1 + (1 + rk )(Xk . the non-random initial wealth.1 -measurable. As before the portfolio process is f consists of X 0. k Sk ) k = 0 : : : n . . 1: f P Then the following processes are martingales under I : 1 S n M k k k=0 and 1 X n M k k k=0 and the following processes are martingales under I : P f k Sk gn=0 k We thus have the following pricing formulas: and f k Xk gn=0: k Simple European derivative security with payoff Ck at time k: f C Vj = Mj IE Mk F j k 1 IE C jF ] = k k j j American derivative security fGk gn=0: k f G Vj = Mj sup IE M F j Tj 1 sup IE G jF ] : = j 2 j 2 Tj The usual hedging portfolio formulas still work.CHAPTER 9. .

Z A X dQ = Z Z dIP 8A 2 G : This shows that Z has the “partial averaging” property. it is the conditional expectation (under the probability measure Q) of X given G . define two probability measures IP (A) = Q(A) 8A 2 G f IP (A) = Z Z Whenever Y is a G -measurable random variable. so I (A) = 0.5 Another Applicaton of the Radon-Nikodym Theorem Let ( F Q) be a probability space. The Radon-Nikodym theorem implies that P is absolutely continuous with respect to the measure I there exists a G -measurable random variable Z such that 1 f IP (A) = 4 Z A Z dIP 8A 2 G A i.. If A 2 G and IP (A) = 0.e. On G . and since Z is G -measurable. In other words. and let X be a non-negative R X dQ = 1.118 9. this is just the definition of IP . We construct the conditional expectation (under Q) of X random variable with given G . Let G be a sub.-algebra of F . the measure I P P f. . we have A XdQ 8A 2 G : Y dIP = Z Y dQ if Y = A for some A 2 G . and the rest follows from the “standard f f machine”. The existence of conditional expectations is a consequence of the Radon-Nikodym theorem. then Q(A) = 0.

2 If is any random variable satisfying (BC).e. Remarks 10. (BC) stands for “Budget Constraint”. i. just regard as a simple European derivative security paying off at time n.1 and 10..1 An Optimization Problem Consider an agent who has initial wealth X 0 and wants to invest in the stock and money markets so as to maximize IE log Xn : Remark 10.Chapter 10 Capital Asset Pricing 10.1 Regardless of the portfolio used by the agent. Remark 10. and starting from this value. there is a hedging portfolio which produces X n = . (BC ) IE n = X0 then there is a portfolio which starts with initial wealth X 0 and produces Xn = at time n.2 show that the optimal obtained by solving the following Constrained Optimization Problem: Find a random variable which solves: Maximize Subject to Equivalently. we wish to Maximize Xn for the capital asset pricing problem can be IE log n IE = X0: X ! 2 (log (! )) IP (! ) 119 . Then X 0 is its value at time 0. f k Xk g1 is a martingale under I . so P k=0 IE nXn = X0 Here. To see this.

Xo = 0: Theorem 1.1’) implies xk = Plugging this into (1. X0 = 0: There are 2n sequences ! in .1) @ f (x : : : x ) = @ g(x : : : x ) k = 1 : : : m m m @xk 1 @xk 1 and g(x1 : : : xm) = 0: For our problem.2’) we get n 1 : n (! k ) 2 1 X IP (! ) = X =) 1 = X : 0 k 0 k=1 . Call them ! 1 !2 : : : !2n .120 Subject to X ! 2 n (! ) (!)IP (! ) .2) xk IP (! k ) = 2n X 1 n (! k )IP (! k ) n (! k )xk IP (! k ) k = 1 : : : 2n = X0 : (1:2 ) 0 (1:1 ) 0 k=1 Equation (1.1) and (1.2) become (1. (1.30 (Lagrange Multiplier) If (x1 Maxmize Subject to then there is a number such that : : : xm) solve the problem f (x1 : : : xm ) g (x1 : : : xm) = 0 (1. Adopt the notation x1 = (! 1) x2 = (!2 ) : : : x2n = (!2n ): We can thus restate the problem as: Maximize 2n X k=1 (log xk )IP (! k ) Subject to In order to solve this problem we use: 2n X k=1 n (! k )xk IP (! k ) .

i.e. 1: n Taking expectations. > 0 and define We maximize f over x > 0: 0 f (x) = log x . Capital Asset Pricing Therefore. (1. xZ: 1 1 f (x) = x . Z = 0 () x = Z The function f is maximized at x 1 f (x) = .CHAPTER 10. X IE ( n ) IE log . n X0 log X0 . 121 0 xk = X! ) k = 1 : : : 2n: ( k n solves the problem Maximize Subject to Thus we have shown that if IE log IE ( n ) = X0 = X0 : n (1. xZ f (x ) = log Z .3). 1 8x > 0 8Z > 0: Let be any random variable satisfying 1 = Z .4).4) Theorem 1.3) then (1..5) we have log . we have 1 IE log . x2 < 0 8x 2 IR: 00 1 log x . 1 0 and so IE log IE log : .5) IE ( n ) = X0 and let = X0 : n From (1.31 If Proof: Fix Z is given by (1. then solves the problem (1.

i. k+1(!1 X0: : : k+1 (!1 : : : !k H ) Sk+1 (!1 : : : !k H ) .. Since f k Xk gn=0 is a martingale under I .e. capital asset pricing works as follows: Consider an agent who has initial wealth and wants to invest in the stock and money market so as to maximize X0 IE log Xn : The optimal Xn is Xn 0 = Xn . Sk+1 (!1 : : : !k T ) : !k T ) . we have P k n Xn = X0: k Xk so = IE nXn jF k ] = X0 k = 0 : : : n Xk = X0 k k (!1 and the optimal portfolio is given by : : : !k ) = X0 .122 In summary.

. 4 -algebra of subsets of . X 1(B) 8B 2 B(IR) where the probabiliy on the right is defined since X 1(B ) 2 F .1 Law of a Random Variable Thus far we have considered only random variables whose domain and range are discrete. X is often called the Law of X – X (B ) = IP .. Thus any random variable X induces a measure X on the measurable space by (IR B(IR)) defined in Williams’ book this is denoted by L X . 1 is a function from B(IR) to F (See Fig. We now consider a general random variable X : !IR defined on the probability space ( F P). Recall that: F is a I is a probability measure on F . B 2 B(IR) (the -algebra of fX 2 Bg = X 1(B) = f! X (!) 2 Bg 2 F i. X 11.e. .1) : !IR is a random variable if and only if X . the set R 4 . P A function X : !IR is a random variable if and only if for every Borel subsets of I ). i.2 Density of a Random Variable The density of X (if it exists) is a function f X X (B ) = Z : IR! 0 1) such that B fX (x) dx 8B 2 B (IR): 123 . 11.Chapter 11 General Random Variables 11. IP (A) is defined for every A 2 F .e.

then d X (x) = fX (x)dx IP fX 2 Bg = 0: 11. We then write where the integral is with respect to the Lebesgue measure on I . Now use the “standard machine” to get the equations for general h. which means that whenever B 2 B(IR) has Lebesgue measure zero.3 Expectation Theorem 3.32 (Expectation of a function of X ) Let h : IR!IR be given. Then IEh(X ) = 4 Z Z h(X (!)) dIP (!) h(x) d X (x) h(x)fX (x) dx: IR. If h(x) = B (x) for some B 4 1 IE 1B (X ) = P fX 2 B g = Z X (B ) = fX (x) dx B which are true by definition. then these equations are = = ZIR IR Proof: (Sketch).1: Illustrating a real-valued random variable X . fX is the Radon-Nikodym derivaR tive of X with respect to the Lebesgue measure.124 X R B {X ε B} Ω Figure 11. Thus X has a density if and only if X is absolutely continuous with respect to Lebesgue measure. .

Then measure on B(IR2 ) (see Fig.CHAPTER 11.2: Two real-valued random variables X Y.2) called the joint law of (X Y ). IEk(X Y ) = 4 Z = = ZZ IR ZZ 2 k(X (!) Y (!)) dIP (!) k(x y ) d X Y (x y) IR k(x y )fX Y (x y ) dxdy 2 . 11.Y) C y { (X. defined by XY induce a X Y (C ) = IP f(X 4 Y ) 2 C g 8C 2 B(IR2 ): The joint density of (X Y ) is a function fX Y : IR2! 0 1) that satisfies X Y (C ) = ZZ C fX Y (x y ) dxdy 8C 2 B(IR2): Y in a manner analogous to the univariate case: We compute the expectation of a function of X fX Y is the Radon-Nikodym derivative of X Y with respect to the Lebesgue measure (area) on IR2 .4 Two random variables Let X Y be two random variables !IR defined on the space ( F P). General Random Variables 125 (X.Y)ε C} Ω x Figure 11. 11.

6 Conditional Expectation Suppose (X Y ) has joint density f X Y .5 Marginal Density Suppose (X Y ) has joint density f X Y .126 11.. Recall that IE h(X )jY ] = IE h(X )j (Y )] depends on ! through Y . there is a function g (y ) (g depending on h) such that 4 IE h(X )jY ](!) = g (Y (!)): We can characterize g using partial averaging: Recall that A 2 (Y )()A = B 2 B(IR).4) . Then the following are equivalent characterizations of g : How do we determine g ? fY 2 Bg for some (6. fY (y ) is the (marginal) density for Y .3) Z B Z Z B IR h(x)fX Y (x y ) dxdy 8B 2 B(IR): (6. Let B IR be given. IR 11.1) Z A g (Y ) dIP = Z A h(X ) dIP 8A 2 (Y ) Z Z IR 1B (Y )g(Y ) dIP = ZZ IR2 Z 1B (Y )h(X ) dIP 8B 2 B(IR) 1B (y)h(x) d X Y (x (6.2) 1B (y)g(y) Y (dy) = g(y )fY (y ) dy = y ) 8B 2 B(IR) (6. i. Then Y (B ) = IP fY 2 B g = IP f(X Y ) 2 IR B g = Z X Z (IR B ) Y = fX Y (x y ) dxdy = ZB Z B 4 fY (y ) dy fX Y (x y ) dx: IR where fY (y ) = Therefore.e. Let h : IR!IR be given.

4): For B Z Z B | IR h(x)fX Y (xjy) dx fY (y ) dy = g (y ) Z Z 2 B(IR) {z j } B IR h(x)fX Y (x y) dxdy: Notation 11.1) becomes IE h(X )jY = y ] = g (y ) = Z IR h(x)fX Y (xjy ) dx: j In conclusion. 2(1 . but fX jY does not.1) (Here g is the function satisfying and g depends on h.1) satisfies (6.1 Let g be the function satisfying IE h(X )jY ] = g (Y ): The function g is often written as g(y ) = IE h(X )jY = y ] and (7.) Theorem 7. first compute Z and then replace the dummy variable y by the random variable Y : IR h(x)fX Y (xjy ) dx j IE h(X )jY ](!) = g (Y (!)): Example 11. 2 exp 1 .1 < < 1.2) Proof: Just verify that g defined by (7.33 If (X IE h(X )jY ] = g (Y ) Y ) has a joint density fX Y . to determine IE h(X )jY ] (a function of ! ). 2 x y + y2 2 2 1 2 . General Random Variables 127 11.CHAPTER 11.1 (Jointly normal random variables) Given parameters: (X Y ) have the joint density 1 >0 1 2 2 > 0 . then fX Y (xjy ) = fXfY (x)y) : Y (y j (7. 2 ) x2 .7 Conditional Density A function fX jY (xjy ) : IR2 ! 0 any function h : IR!IR: 1) is called a conditional density for X given Y provided that for g (y ) = Z IR h(x)fX Y (xjy ) dx: j (7. : Let fX Y (x y) = 2 1 2 1 p1 .

2(1. 2(1 . 1 2 y 2 dx:e 1 .1 e . 1 2 dx y .1 = (1 . 2 ) 2 1 2 2 2 2 1 = ..128 The exponent is . 2 1 : fX Y (x y) = 2 1 2 1 p 1.1 p 1 Thus Y is normal with mean 0 and variance 2 . 2 Conditional density. 2) 12 x . e 2 1 . 1 Therefore. 2(1 . From the expressions . 2 ) 12 x . = p1 e 2 2 . 1 y . 2 . 2 ) 1 x 2 1 2 .1 2) 2 2 2 2 1 x. du = p1. 2 y = 2 2 2 2 1 2 1. 2 y2 2 2 using the substitution u = 1 y2 2 2 2 . 2 1 y2 2 2 1y 2 fY (y) = p 1 e. 1y 2 2 e. 21 y : 1 = p 2 1 1. 2 2 2 2 we have fX jY (xjy) = fXfY (x y) Y (y) 2 1 1 p 1 e. 2 x y + y2 2 2 2(1 . 2 ) 2 : 1 x. 2 ) Z1 IE X jY = y] = xf (xjy) dx = 1 y 2 . 2 In the x-variable. 2(1.2 1 1 x. 1 2 = Z1 y 2 Y =y 1 2 # . 2(1. x2 . y 2 + y2 (1 . y fX jY (xjy) dx 2 . " X. fX jY (xjy) is a normal density with mean 21 y and variance (1 . 2 ) 1 " 1 2 2 # x .1 X jY IE 2. 1 y 2 : 2 2 1 2 1 p 1 We can compute the Marginal density of Y as follows fY (y) = Z1 1 1 Z 1 e. u2 du:e. 2 ) 1 = .

. 2 ) 2: 1 # (7. ):(X . 1 2 Y 2 # = (1 . )T:A:(X . j )2 is the variance of Xj .4) yields IEX = IE 1 2 IEY = 0 (7. 2 2 1 2 ::: 2) n where 2 j = IE (Xj .3) and (7.3) " X. The random variables in if and only if A. the (i j )th element of A. No other estimator based on Y can have a smaller expected square error 1 (Homework problem 2. A 1 is the covariance matrix 1 ::: . the best estimator of X is 21 Y . i.8 Multivariate Normal Distribution Please see Oksendal Appendix A.e. General Random Variables From the above two formulas we have the formulas 129 IE X jY ] = IE 1 2 Y (7. X are independent A 1 = diag( .1 is diagonal. 2 ) 2.4) Taking expectations in (7. ) : 2 ) =2 =( 4 Here. .e.5) " X.CHAPTER 11.1 is IE (Xi . 2 ) 2: 1 (7. 4 . This estimator is unbiased (has expected error zero) and the expected square error is (1 . i )(Xj . Let denote the column vector of random variables (X1 X2 : : : Xn )T . 1 2 Y 2 Y = (1 . and A is an n T T n ) = IE X = (IEX1 : : : IEXn) n nonsingular matrix.1). )T h i i. and the corresponding has a multivariate normal distribution if and only if column vector of values (x1 x2 : : : xn )T . 11.6) Based on Y . 1 (X . A 1 = IE (X . the random variables have the joint density X X x p o n fX(x) = (2detnA exp . j ).

1 . 3 5 det A = 1 2 1. . The random variables are jointly normal and independent if and only if for any real column vector = (u 1 : : : un )T u 8n 9 8n 9 <X = <X 1 2 2 = IEeuT :X = IE exp : uj Xj = exp : 2 j uj + uj j ] : j =1 j =1 4 .9 Bivariate normal distribution Take n = 2 in the above definitions. and let 4 = IE (X1 . #) ( " 1 (x1 . 1 )2 . Thus. If any n random variables X1 X2 : : : Xn have this moment generating function.1 and mean vector . adjusted to account for the possibly non-zero expectations: fX1 X2 (x1 x2) = 2 1 1 p 2 1. 2(1 . 2 ) : 1 2 . 2 A=4 A = 1 " . 2 (x1 .130 11.10 MGF of jointly normal random variables Let = (u1 u2 : : : un )T denote a column vector with components in IR. 1 p 1 2 2 2 (1 ) 1 (1 2 ) 2 . then they are jointly normal. 2 ) + (x2 . 2 and we have the formula from Example 11. and let have a multivariate normal distribution with covariance matrix A . . 1 )(X2 .1. and we can read out the means and covariances. 2 1 1 2 1 2 2 2 # . p 1 (1 2 ) 2) 1 2 (1 2 1 . 1 )(x2 . Then the moment generating function is given by u X T IEeu :X = T eu :XfX1 X2 : : : Xn (x1 x2 : : : xn ) dx1 : : :dxn o n = exp 1 uT A 1 u + uT : 2 1 . 2) 1 2 1 2 11.1 Z ::: Z 1 . 2 )2 : 2 2 2 exp .

then we know B1 B2 : : : Bk . u 8n 9 8n 9 <X = <X = TY IEeu = IE exp : uj Yj = exp : 1 u2 : 2 j j =1 j =1 B0 = 0 Bk = k X F P).34 Proof: fBk gn=0 is a martingale (under I ). Conversely. Define the filtration j =1 Yj k = 1 : : : n: B2 : : : Bk . standard normal random variables defined on ( j where I is the market measure. As before we denote the column vector (Y1 : : : Yn )T by P therefore have for any real colum vector = (u1 : : : un )T . 12.1 Discrete-time Brownian Motion Let fYj gn=1 be a collection of independent. P k IE Bk+1 jF k ] = IE Yk+1 + Bk jF k ] = IEYk+1 + Bk = Bk : 131 . We Define the discrete-time Brownian motion (See Fig. B1 : : : Yk = Bk .Chapter 12 Semi-Continuous Models 12. F0 = f g F k = (Y1 Y2 : : : Yk ) = (B1 B2 : : : Bk ) k = 1 : : : n: Theorem 1.1 . Bk. Y.1): If we know Y1 Y2 : : : Yk . if we know B1 then we know Y1 = B1 Y2 = B2 .

2 The Stock Price Process Given parameters: > 0. 1 y2 2 dy: Then which is a function of B k alone.132 B k Y2 Y1 0 1 k 2 Y 3 3 Y 4 4 Figure 12. Define . 2 2 ) n o . 1 2 )k 2 Note that n o k = 0 : : : n: 1 Sk+1 = Sk exp Yk+1 + ( . S0 > 0. the volatility. 2 IR. IE h(Yk+1 + Bk )jF k ] = g(Bk ) 12. The stock price process is then given by Sk = S0 exp Bk + ( .1 h(y + b)e . k IE h(Bk+1 )jF k ] = IE h(Yk+1 + Bk )jF k ]: 1 Z g(b) = IEh(Yk+1 + b) = p 2 1 Proof: Note that Use the Independence Lemma.1: Discrete-time Brownian motion. Theorem 1.35 fBk gn=0 is a Markov process. the mean rate of return. the initial stock price.

If P . P P k=0 F ).CHAPTER 12.3 Remainder of the Market The other processes in the market are defined as follows.4 Risk-Neutral Measure k Sk+1 Sk Xk Mk+1 . Wealth process: X0 given. Mk + Mk : f Definition 12. ::: n 1 . 2 2) = 2 : k and 12. IE Sk+1jF k ] = Sk IE e Yk+1 jF k ]:e 1 2 1 2 = Sk e 2 e 2 = e Sk : . we say that I is a risk-neutral measure.1 Let I be a probability measure on ( P n o Sk Mk n f f is a martingale under I . Semi-Continuous Models 133 . Money market process: Portfolio process: Mk = erk k = 0 1 : : : n: 0 Each 1 k is F k -measurable. equivalent to the market measure I . 1 2 2 Thus k +1 = log IE SkS jF k ] = log IE SS+1 F k k k k 1 var log SS+1 = var Yk+1 + ( . nonrandom. Discounted wealth process: r k Sk+1 + e (Xk . e Sk ) + e Xk Xk+1 Mk+1 = 12. k Sk ) r r k (Sk+1 . Xk+1 = = Each Xk is F k -measurable.

P Theorem 6. Hedging a short position: Sell the simple European derivative security V n . .6 Arbitrage Definition 12. fX f V X0 = IE Mn = IE Mnn : n Remark 12. then Construct a portfolio process 0 . 1 which starts with X 0 and ends with Xn = Vn . Receive X0 at time 0.134 f Theorem 4. and say it is F n -measurable.5 Risk-Neutral Pricing Let Vn be the payoff at time n. regardless of the portfolio process used to generate it.37 (Fundamental Theorem of Asset Pricing: Easy part) If there is a risk-neutral measure.36 If I is a risk-neutral measure.1 Hedging in this “semi-continuous” model is usually not possible because there are not enough trading dates. 12. P Proof: n Xk on Mk k=0 is f X IE Mk+1 F k k+1 f = IE = X = Mk : k k Sk+1 Sk Xk Mk+1 . ::: n f P If there is a risk-neutral measure I . Mk + Mk F k S X f S IE Mkk+1 F k . Note that Vn may be path-dependent. Mkk + Mk +1 k k 12. then every discounted wealth process P f a martingale under I . This difficulty will disappear when we go to the fully continuous model.2 An arbitrage is a portfolio which starts with X 0 = 0 and ends with Xn satisfying IP (Xn 0) = 1 IP (Xn > 0) > 0: (I here is the market measure). then there is no arbitrage.

= 0) = 1. let X0 = 0. 1 2g 2 2 Sk : = e r : Mk = r. r . 1 2)o 2 Mk+1 Mk If S S IE Mkk+1 F k = Mkk :IE exp f Yk+1 gjF k ] : expf . r . and let Xn be the final wealth corresponding P Proof: Let I be a risk-neutraln on f is a martingale under I .1) and (6. We have f f IP (Xn = 0) = 1 =) IP (Xn > 0) = 0 =) IP (Xn > 0) = 0: 12. o n Sk = S0 exp Bk + ( .CHAPTER 12. n o Sk+1 = Sk : exp n Yk+1 + ( . 1 2)k . Semi-Continuous Models 135 f measure.2) imply I (Xn P This is not an arbitrage. 2 Therefore.1) 0) = 1.2) f f IP (Xn 0) = 1 =) IP (Xn < 0) = 0 =) IP (Xn < 0) = 0 =) IP (Xn 0) = 1: f (6.7 Stalking the Risk-Neutral Measure Recall that Y1 Y2 : : : Yn are independent. 2 Sk+1 = S0 exp (Bk + Yk+1 ) + ( . . we must seek further. 1 2)(k + 1) o 2 n 1 2) : = Sk exp Yk+1 + ( . standard normal random variables on some probability space ( F P). P to any portfolio process. If 6= r. . We have (6. the market measure is risk neutral. r . 1 2g 2 +1 S = Mkk : expf 1 2 g: expf . Since Xk Mk k=0 fX fX IE Mn = IE M00 = 0: n Suppose IP (Xn (6.

1 2 2 n ~ o S = Mkk : exp Yk+1 . 2 2 )5 : j =1 Cameron-Martin-Girsanov’s Idea: Define the random variable Properties of Z : Z 0. 1 2 )o 2 Mk+1 Mk o n S = Mkk : exp (Yk+1 + r ) . P f IP ( ) = IEZ = 1: . 1 2g 2 S = Mkk : 3 2n X 1 Z = exp 4 (. 9 8n = <X IEZ = IE exp : (. n 2 j =1 = exp n 2 2 2 : exp . Yj ) : exp . ~ Yk+1 = Yk+1 + . 1 2 2 .136 Sk+1 = Sk : exp n Yk+1 + ( . I is a probability measure. f ~ P We want a probability measure I under which Y1 dom variables. r: ~ : : : Yn are independent. 2 2g +1 S 1 = Mkk : expf 2 2 g: expf. where The quantity . standard normal ran- i S fh 1 f S ~ IE Mkk+1 F k = Mkk :IE expf Yk+1 gjF k : expf. Then we would have r is denoted and is called the market price of risk. r . n 2 2 = 1: Define f IP (A) = 0 for all A 2 F and Z A Z dIP 8A 2 F : f Then I (A) P f In other words. Yj .

uj + 1 2 ) + (uj . For this. f ~ ~ Compute the moment generating function of (Y1 : : : Yn ) under I : P 2n 3 2n 3 n X ~5 X X f IE exp 4 uj Yj = IE exp 4 uj (Yj + ) + (. 2 2 )5 j =1 2j=1 3 n X 1 2 = exp 4 ( 2 uj . standard normal under I . 1 2)5 2 j =1 j =1 j =1 3 2n 3 2n X X = IE exp 4 (uj . P Verification: Y1 Y2 : : : Yn : Independent.CHAPTER 12. 1 2 )5 2 j =1 j =1 2n 3 2n 3 X1 X 1 = exp 4 2 (uj . )Yj 5 : exp 4 (uj . Yj . 1 2) j 2 ~ ~ Y = Y1 + : : : Yn = Yn + : Z > 0 almost surely. f IP (A) = Z A Z dIP 8A 2 F f IEX = IE (XZ ) for every random variable X . standard normal under I . Yj . it suffices to show that ~ ~ Y 1 = Y1 + : : : Yn = Yn + f are independent. Semi-Continuous Models 137 f P We show that I is a risk-neutral measure. and P 2n 3 2n 3 X 5 X IE exp 4 uj Yj = exp 4 1 u2 5 : 2 j j =1 j =1 i h Z = exp Pn=1 (. 1 2) 5 2 2 j =1 2n 3 X = exp 4 1 u2 5 : 2 j j =1 . )25 : exp 4 (uj .

( . f ~j since j =1 1 j =1 e . 1 2 )n 2 j =1 8 n 9 < X = = S0 exp : (Yj + r ) . 1 2)n 2 o n 8 n 9 < X = = S0 exp : Yj + ( . Payoff at time n is (Sn . Price at time zero is 2 0 1+3 n (Sn .138 12. It does not depend on .2under IP . K : p 1 e 2n2 db 2 n Pn Y is normal with mean 0. 1 2 )n . 1 2)n 2 j =1 8 n 9 < X = 1 ~ = S0 exp : Yj + (r . 8 n 9 n = Z . r)n + ( .8 Pricing a European Call Stock price at time n is Sn = S0 exp Bn + ( . K A 5 f f ~j IE M 0 2 : . rn o + . K )+ . variance n. K )+ = IE 4e rn @S exp < X Y + (r . b2 This is the Black-Scholes price. 1 2)n= . .1 S0 exp b + (r . 2 2)n : j =1 .

as 139 k!1: .38 (Law of Large Numbers:) 1 M !0 k k almost surely.1 Symmetric Random Walk Toss a fair coin infinitely many times. Define Xj (!) = 1 Set ( .2 The Law of Large Numbers We will use the method of moment generating functions to derive the Law of Large Numbers: Theorem 2.Chapter 13 Brownian Motion 13.1 if if !j = H ! j = T: M0 = 0 Mk = k X j =1 Xj k 1: 13.

as k!1: u 'k (u) = IE exp p Mk k u u 1 = 2 e pk + 1 e pk 2 . Then k . Theorem 3. (Def. of Mk :) (Independence of the Xj ’s) 1 u 2e k + 1e 2 u .140 Proof: 'k (u) = IE exp u Mk 9 8kk <X u = = IE exp : k Xj j =1 k Y = IE exp u Xj k j =1 = which implies. u k k! lim log 'k (u) = xlim0 ! 1 log 1 eux + 1 e ux 2 2 . k . x (L’Hˆ pital’s Rule) o Therefore. u eux .f.g. k! lim 'k (u) = e0 = 1 1 which is the m. 13. for the constant 0. u k k log 'k (u) = k log 1 e k + 1 e 2 2 Let x = 1 . .39 (Central Limit Theorem) 1 p Mk ! k Proof: Standard normal. = xlim0 2 ux ! 1e + 2 = 0: u e ux 2 1 e ux 2 .3 Central Limit Theorem We use the method of moment generating functions to prove the Central Limit Theorem.

x2 u eux .g. . 13.4 Brownian Motion as a Limit of Random Walks Let n be a positive integer. u e ux = xlim0 2 1 ux 2 1 ux ! 2x 2 e + 2 e . then set 1 1 B (n) (t) = pn Mtn = pn Mk : k 0 is not of the form n . (L’Hˆ pital’s Rule) o = xlim0 2 ! = xlim0 ! 1 u2 : =2 Therefore. u e ux = xlim0 1 ux 1 1 ux : xlim0 2 2x2 ! ! 2e + 2e . 13.f. If t k 0 is of the form n . then define B(n) (t) by linear interpolation (See Fig. 141 1 log 'k (u) = k log 1 e pk + 2 e 2 1 Let x = pk . u2 e ux 2 2 . u pk : lim log 'k (u) = xlim0 ! k! 1 u eux .1). u eux . u2 eux . 2x 2 (L’Hˆ pital’s Rule) o lim ' (u) = e 2 u k! k 1 1 2 which is the m. for a standard normal random variable. Here are some properties of B(100)(t): If t . u e ux 2 . Then u .CHAPTER 13. Brownian Motion so that. . 1 log 2 eux + 1 e ux 2 .

142 k/n (k+1)/n Figure 13. Properties of B(100)(1) : 1 B (100)(1) = 10 100 X 1 IEB(100)(1) = 10 j =1 100 X Xj IEXj = 0: var(Xj ) = 1 (Approximately normal) var(B (100)(1)) = 100 Properties of j =1 100 1 X j =1 B(100)(2) : 1 B (100)(2) = 10 200 X IEB(100)(2) = 0: var(B (100)(2)) = 2: Also note that: j =1 Xj (Approximately normal) B (100)(1) and B (100)(2) .1: Linear Interpolation to define B (n) (t). Chapter 2. B (100)(t) is a continuous function of t. let n!1 in B (n) (t) t 0.5 Brownian Motion (Please refer to Oksendal.) . To get Brownian motion. 13. B (100)(1) are independent.

13. IEB (s) = 0 var(B (s)) = s IEB(t) = 0 var(B (t)) = t IEB(s)B (t) = IEB(s) (B(t) . B has independent. A random variable properties: 1. 3. B (s)) + IEB 2 (s) (t } | {z } | = s: 0 s . then Y1 Y2 : : : Yn are independent. B (t) is a continuous function of t. F. tj 1 8j: . so B (s) and B (t) = (B (t) . t B(t) (see Fig. Brownian Motion 143 B(t) = B(t. IEYj = 0 8j var(Yj ) = tj . normally distributed increments: If 0 = t0 < t1 < t2 < : : : < tn and Y1 = B (t1 ) .2) is called a Brownian Motion if it satisfies the following B (0) = 0.CHAPTER 13. Moreover. B(s)) + B(s)] = IEB (s)(B{z) . B(tn 1 ) .2: Continuous-time Brownian Motion. B (s)) + B (s) are jointly normal. B (t0 ) Y2 = B(t2 ) .ω) ω (Ω. 2. B(t1 ) : : : Yn = B (tn ) . 13.P) Figure 13. B (s) are independent.6 Covariance of Brownian Motion Let 0 s t be given. Then B (s) and B (t) .

required by the -algebra properties. fF (t)gt 0 is called the filtration generated by the Brownian motion. Do this for all possible numbers s 2 0 t] and C 2 B(IR). B (tn 1 ) are independent of F (t). Then 0 < t1 < t2 < : : : < tn (B (t1 ) B (t2) : : : B (tn )) is jointly normal with covariance matrix 2 3 IEB 2(t1 ) IEB (t1 )B (t2) : : : IEB(t1 )B (tn ) 6 7 2 C = 6:IEB:(t:2:):B:(:t1 ) : : : : :IEB: :(t2:): : : : : :: :: :: : : IEB:(:t:2:):B (t:n: )7 6 ::: : :: 4 ::: : ::: : : :7 5 IEB(tn)B (t1 ) IEB(tn )B(t2 ) : : : IEB2 (tn) 2 3 t1 t1 : : : t1 6 7 = 6t1 : : t:2: : ::::::: : : t:2:7 6: : 7 4 5 t1 t2 : : : tn 13.8 Filtration generated by a Brownian Motion fF (t)gt Required properties: For each t. B(t) B(t2 ) . Let s 2 0 t] and C 2 B(IR) be given. we have IEB(s)B(t) = s ^ t: 13. Here is one way to construct F (t). Put the set Then put in every other set fB(s) 2 C g = f! : B(s !) 2 C g in F (t). B(t1 ) : : : B (tn ) .144 Thus for any s 0. t 0 (not necessarily s t). the Brownian motion increments B (t1 ) . First fix t. B (t) is F (t)-measurable. . For each t and for t < t1 0 < t2 < < tn .7 Finite-Dimensional Distributions of Brownian Motion Let be given. . This F (t) contains exactly the information learned by observing the Brownian motion upto time t.

s) 2 = Z (s): h i 13. Then Z (t) = exp . 1 2 (t . 2 2 (t .9 Martingale Property Theorem 9. Proof: Let 0 s t be given. 1 2((t . s)g 2 n1 2 o = Z (s) exp 2 (. r = 0. B (s)) . ) var(B (t) . Proof: Let 0 n o s t be given. Brownian Motion 145 13. B (s) + B(s)) . B (s)] + B (s) = B (s): Theorem 9.40 Brownian motion is a martingale. B (s)) . (B (t) . B (s)) . ~ 2 p p p . p .CHAPTER 13. Then IE B (t)jF (s)] = IE (B(t) . 1 2 (t . Then 1 = IE Z (s) expf. B(s)) + B(s)jF (s)] = IE B (t) .10 The Limit of a Binomial Model Consider the n’th Binomial model with the following parameters: un = 1 + n : “Up” factor. (B(t) . n : “Down” factor. ~ 2 qn = 1 . dn = 1 . 1 2 t 2 is a martingale. ( > 0). s)g F (s) IE Z (t)jF (s)] = IE expf. pn = u1n ddnn = 2 == nn = 1 . B(t) . . s) + s)g F (s) 2 = Z (s)IE expf. (B (t) .41 Let 2 IR be given.

Proof: Recall that from the Taylor series we have expf B (t) . Mnt ) log(1 . . Mk ): 2 In the n’th model. Note that the correction martingale. and let ] k (T ) denote the number of T in the first k tosses. the distribution of S (n) (t) converges to the distribution of where B is a Brownian motion. Set S0 f P Under I . pn ) 2 n n pn ) . 4 n 2 n 4 n 2 1p 2 n 1 2 pn .4 n . pn ) 2 ! 1 = . Let t = k for some k. take n steps per unit time. 1 log(1 . 1 | {z } | {z } !0 !Bt As n!1. p ) 2 2 = nt 1 log(1 + 2 + Mnt 1 2 + O(n 3=2 ) = nt . . 1 2t. . Then ]k (H ) + ]k (T ) = k ]k (H ) . the price process S (n) is a martingale.42 As n!1. ]k (T ) = Mk which implies. 1 2 tg 2 . 4n ! 1 2 + 1 p + 1 2 + O(n 3=2 ) + Mnt 1 pn . the distribution of log S (n) (t) approaches the distribution of B(t) .146 Let ]k (H ) denote the number of H in the first k tosses. 2 2 t + O (n 2 ) 1 1 1 + pn Mnt + n Mnt O(n 2 ) . 1 2 log(1 + 1 2 pn ) + 1 log(1 . ]k (H ) = 1 (k + Mk ) 2 ]k (T ) = 1 (k . pn . and let n 1 (nt+Mnt ) 1 (nt Mnt ) 2 2 1 . 1 2t is necessary in order to have a 2 log(1 + x) = x . : Theorem 10. 2 . S (n)(t) = 1 + pn (n) = 1. 1 x2 + O(x3) 2 so log S (n)(t) = 1 (nt + Mnt ) log(1 + p ) + 1 (nt .

starting at x 6= 0. Brownian Motion 147 B(t) = B(t. F. then IP x puts all its probability on a completely different set from I .4).3: Continuous-time Brownian Motion. P 13. 13. 13. denote the corresponding probability measure by IP x (See Fig. Proof: Let s 0 2 6 ) IE h(B (s + t)) F (s) = IE 6h( B(s + t{z.ω) x ω t (Ω.) For a Brownian motion B (t) that starts at 0. and for such a Brownian motion we have: IP x (B(0) = x) = 1: Note that: If x 6= 0. 13. B(s) + 4 | } Independent of F (s) F t 0 be given (See Fig. Px) Figure 13.CHAPTER 13.12 Markov Property for Brownian Motion We prove that Theorem 12.11 Starting at Points Other Than 0 (The remaining sections in this chapter were taught Dec 7.43 Brownian motion has the Markov property.3). we have: IP (B(0) = 0) = 1: For a Brownian motion B (t) that starts at x. P The distribution of B (t) under IP x is the same as the distribution of x + B (t) under I . (s)-measurable B{zs) |( } 3 7 ) F (s)7 5 .

Fix x > 0 and define = min ft 0 B(t) = xg : Then we have: IE h( B( + t) ) F ( ) = g(B( )) = IE x h(B(t)): . Use the Independence Lemma. B(s) + x )] 2 6 = IE 6h( x + 4 = IE xh(B (t)): Then same distribution as B (s + t) .5. Example 13. B (s) B (t) |{z} 3 7 )7 5 IE h (B(s + t) ) F (s) = g (B(s)) = E B (s)h(B (t)): In fact Brownian motion has the strong Markov property. 13.1 (Strong Markov Property) See Fig.4: Markov Property of Brownian Motion. Define s+t g (x) = IE h( B (s + t) .148 B(s) s restart Figure 13.

Then exp B (t ^ ) .CHAPTER 13. and 1 IE exp B(t ^ ) .5: Strong Markov Property of Brownian Motion. 1 2 (t ^ ) 2 n o is a martingale. (y .1 IE h(B (s + t)) F (s) = g(B (s)) = IE h(B( + t)) F ( ) = 13. Define Fix Z 1 h(y )p(t B (s) y) dy: Z .1 1 h(y )p(t x y) dy: . Brownian Motion 149 x τ restart Figure 13. 2 2 (t ^ ) = 1: n o .1 = min ft 0 B (t) = xg : > 0.13 Transition Density Let p(t x y ) be the probability that the Brownian motion changes value from x to y in time t. τ+t 13. p(t x y) = p 1 e 2 t g(x) = IE xh(B(t)) = . and let be defined as in the previous section.x)2 2t Z 1 h(y )p(t x y) dy: .14 First Passage Time Fix x > 0.

Brownian motion reaches level x with probability 1. to get IE expf x .IE e Letting .f. p 2 > 0: p (14. We use the Reflection Principle below (see Fig.g.r. px e x 2 . so IP f < 1g = 1 IE expf.6). using the Bounded Convergence Theorem.3) yields . 13. 2 (t ^ ) = : 2 t! if = 1 0 0 expf B (t ^ ) .: IEe Differentiation of (14. We have the m. .2) =1 2 2 .1).t. 1 2 (t ^ )g e x : 2 .150 We have 8 n 12 o <e 1 2 if < 1 lim exp . =e x . y2 2t dy . x: (14. = . The expected time to reach level x is infinite. 1 2 g = e 2 Let . IP f t B (t) < xg = IP fB (t) > xg IP f tg = IP f t B(t) < xg + IP f t B (t) > xg = IP fB (t) > xg + IP fB (t) > xg = 2IP fB (t) > xg Z e = p2 2 tx 1 . 2 : #0. 1 2 Let h 2 g1 f < i 1g = 1: #0 to get IE 1 f < 1g = 1.1) Let t!1 in (14. 1 (14. we obtain IE = 1: (14.3) w.4) Conclusion.

@a g (a(t)): @t @t IEe .6: Reflection Principle in Brownian Motion.CHAPTER 13. Using the substitution z = yt p dz = p dy we get t IP f Density: tg = p2 Z 1 2 x pt e . Laplace transform formula: = e t f (t)dt = e x . . Brownian Motion 151 shadow path x τ t Brownian motion Figure 13. z2 2 dz: x2 2t @ f (t) = @t IP f F (t) = tg = p x 3 e 2 t . Z 0 1 p 2 : . which follows from the fact that if Zb a(t) g (z) dz then @F = .

152 .

even when not explicitly mentioned. are independent. B (t0 ) : : : B(tn ) . Futures and other Derivative Securities. Options.2 First Variation Quadratic variation is a measure of volatility. 14. !IR.Chapter 14 The Itˆ Integral o The following chapters deal with Stochastic Differential Equations in Finance. B(tk )]2 = tk+1 .3. 13. 153 FV (f ). B (t ! ) : 0 1) 1.1995 2.1 Brownian Motion (See Fig. then the increments B (t1) . always in the background. Oksendal. B(tk )] = 0 IE B (tk+1 ) . References: 1. B. B (0) = 0 Technically. function f (t). IP f! B(0 !) = 0g = 1. of a . J. 2. B(tn 1 ) . Prentice Hall. Springer-Verlag. B (t) is a continuous function of t. and IE B (tk+1 ) . If 0 = t0 t1 : : : tn . has the following properties: Brownian motion.) ( F P) is given. 1993. Hull. 3. Stochastic Differential Equations.normal. tk : 14. First we will consider first variation.

f (0)] . f (tk ) = f (tk )(tk+1 .1: Example function f (t). f (tk )j: Suppose f is differentiable. f (t2 )] = f (t) dt + (. first variation measures the total amount of up and down motion of the path. the first variation over the interval 0 T ] is given by: FV 0 T ](f ) = f (t1) . The general definition of first variation is as follows: Definition 14. We then define FV 0 T ](f ) = lim 0 ! jj jj n 1 X .154 f(t) t2 t1 T t Figure 14.f (t)) dt + f (t) dt: 0 0 0 Zt1 0 ZT 0 Zt2 ZT t1 t2 = jf (t)j dt: 0 Thus.1. f (t1 )] + f (T ) . tk ): ::: n .1 (First Variation) Let = ft0 t1 : : : tn g be a partition of 0 T ]. 14. tk ): 0 . For the function pictured in Fig. k=0 jf (tk+1) . there is a point tk such that f (tk+1 ) . 0 = t0 t1 : : : tn = T: The mesh of the partition is defined to be jj jj = k=0max 1(tk+1 . f (t2) .. Then the Mean Value Theorem implies that in each subinterval t k tk+1 ]. i.e.

IP f! 2 hB(: !)i(T ) = T g = 1: In particular. tk ) 0 0 FV 0 T ](f ) = lim 0 ! jj jj n 1 X k=0 jf (tk )j(tk+1 . = lim jj jj jf (t)j2 dt jj !0 ZT 0 k=0 jf (tk )j2(tk+1 . tk ) 0 hf i(T ) jj lim 0 jj jj: lim 0 ! ! jj jj jj jj n 1 X . tk ) 0 0 = 0: Theorem 3. f (tk )j = n 1 X .1 (Quadratic Variation of Differentiable Functions) If f is differentiable. then hf i(T ) = 0.2 (Quadratic Variation) The quadratic variation of a function f on an interval is 0 T] hf i(T ) = lim 0 jf (tk+1 ) . tk )2 0 jj jj: and n 1 X . jf (tk )j(tk+1 . The Itˆ Integral o Then 155 n 1 X . k=0 jf (tk+1) . k=0 jf (tk )j2(tk+1 . k=0 . Remark 14. f (tk )j2 = n 1 X . f (tk )j2: ! k=0 jj jj n 1 X .CHAPTER 14. k=0 and jf (tk+1) .3 Quadratic Variation Definition 14. the paths of Brownian motion are not differentiable. . because n 1 X .44 hBi(T ) = T or more precisely. k=0 jf (tk )j2(tk+1 . tk ) = jf (t)j dt: 0 ZT 0 14.

=2 n 1 X . tk ) = B(tk+1 ) . (tk+1 .T = We want to show that jj jj n 1 X . T ) = = = n 1 X . (tj+1 . B(tk ). 2(tk+1 . (tk+1 . tk )2 n 1 X . k=0 2 Dk . the terms var(Q . (if X is normal with mean 0 and variance 2. k=0 2 Dk : Q . T ) = 0: !0 Consider an individual summand 2 Dk . tk ) Dj2 . tk )] = 0: 2 Dk . tk )2] k=0 n 1 X . tk )Dk + (tk+1 . tk ) IE (Q . Define the sample quadratic variation 0 T ]. 2 var Dk . then IE (X 4) = 3 4) k=0 2jj jj = 2jj jj T: Thus we have k=0 (tk+1 .156 Proof: (Outline) Let = ft0 t1 : : : tn g be a partition of B (tk+1 ) . tk )] 4 2 IE Dk . (tk+1 . T ) 2jj jj:T: . tj ) and are independent. tk )2 ] (tk+1 . so IE (Q . set D k = Q = Then n 1 X . (tk+1 . T ) = IE For j n 1 X . tk )2 + (tk+1 . tk ): This has expectation 0. k=0 2 Dk . 6= k. (tk+1 . tk )2 . T ) = 0 var(Q . tk )]: lim (Q . 2(tk+1 . B(tk )]2 . To simplify notation. so k=0 n 1 X . (tk+1 . k=0 3(tk+1 .

B (tk ))2 ' tk+1 . Furthermore. Brownian motion has absolute volatility 1. . T1 T2 . tk )] = 2(tk+1 . B(tk )) : 2 and compute the squared sample absolute volatility 1 n 1 X . tk )2 : When (tk+1 .4 Quadratic Variation as Absolute Volatility On any time interval T1 T2]. (tk+1 . and we have the approximate equation (B (tk+1 ) . tk )] = 0: We showed above that var (B (tk+1 ) . consider the equation hBi(T ) = T = 1 dt 0 ZT 8T 0: This says that quadratic variation for Brownian motion accumulates at rate 1 at all times along almost every path. tk ) is small. this approximation becomes exact. tk which we can write informally as dB(t) dB(t) = dt: 14. B (tk ))2 . The Itˆ Integral o As jj 157 jj!0. (tk+1 . T1 k=0(B(tk+1 ) . so jj lim (Q . This is approximately equal to hBi(T2) .CHAPTER 14.2 (Differential Representation) We know that IE (B(tk+1 ) . T1 1 As we increase the number of sample points. var(Q . we can sample the Brownian motion at times T1 = t0 t1 : : : tn = T2 T2 . T ) = 0: jj !0 Remark 14. tk )2 is very small. T1 = 1: T2 . In other words. (tk+1 . B(tk ))2 . hBi(T1)] = T2 . T )!0.

B (t) is F (t)-measurable. then Remark 14. and the s t=) every set in F (s) is also in F (t).e. the increments B (t1 ) . (t) t 0. The functions B (t) and (tk ) can be interpreted as follows: Think of B (t) as the price per unit share of an asset at time t. because the paths of Brownian motion are not differentiable.r..e. 3. 2. We call such a elementary process. is square-integrable: IE We want to define the Itˆ Integral: o ZT 0 2(t) dt < 1 8T: I (t) = Zt 0 (u) dB (u) t 0: f (t) is a differentiable function. i. .t. 8t. B (t) B (t2 ) . The integrand is 1. B(t) t 0. For t t1 : : : tn . 14.2). 14.. a differentiable function) If we can define Zt 0 (u) df (u) = Zt 0 (u)f (u) du: 0 This won’t work when the integrator is Brownian motion. 0 = t0 t1 : : : tn = T: an Assume that (t) is constant on each subinterval t k tk+1 ] (see Fig. B (tn 1 ) are independent of F (t). with associated filtration F (t) t 0. where is adapted) (t) is F (t)-measurable 8t (i.3 (Integral w.158 14. B (t1 ) : : : B (tn ) . . 2.5 Construction of the Itˆ Integral o The integrator is Brownian motion following properties: 1.6 Itˆ integral of an elementary integrand o Let = ft0 t1 : : : tn g be a partition of 0 T ].

B (tk )]: ˆ 14. B(t0)] + (t1) B(t2) . if tk j =0 (tj ) B (tj +1 ) . this gain is given by: o : : : tn as the trading dates for the asset. B(t1 )] + (t2) B(t) . Think of t0 t1 Think of (tk ) as the number of shares of the asset acquired at trading date t k and held until trading date tk+1 . B (tj )] + (tk ) B (t) . | (t0) ] B{z } 0 t t1 > < =B (0)=0 I (t) = > (t0 ) B (t1) . The Itˆ Integral o 159 δ( t ) = δ( t 1 ) δ( t ) = δ( t ) 0 δ( t )= δ( t 3 ) 0=t0 t1 t2 t3 t4 = T δ( t ) = δ( t 2 ) Figure 14. Then the Itˆ integral I (t) can be interpreted as the gain from trading at time t.7 Properties of the Ito integral of an elementary process Adaptedness For each t Linearity If I (t) is F (t)-measurable. In general. B (t1 )] t t t2 > : (t0) B(t1) . I (t) = Zt 0 (u) dB (u) J (t) = ( (u) Zt 0 (u) dB (u) then I (t) J (t) = Zt 0 (u)) dB (u) . 8 > (t0) B(t) . B (t0 )] + (t1 ) B (t) .CHAPTER 14. I (t) = k 1 X .2: An elementary function . B(t2)] t1 t t3: 2 t tk+1 .

3: Showing s and t in different partitions. We prove the martingale property for the elementary process case. Theorem 7. + j =`+1 (tj ) B (tj +1 ) . B (tj )] + (tk ) B (t) . B(t` )) F (s) = (t` ) (IE B(t`+1)jF (s)] . B(tj )) F (s)5 = (tj )(B (tj +1 ) . B (t` )] . .160 s t l t l+1 t . i. t k t k+1 Figure 14. We treat the more difficult case that s and t are in different Proof: Let 0 subintervals. B (tj )] + (t` ) B (t`+1 ) .45 (Martingale Property) I (t) = is a martingale. B (tj )): . k 1 X . 14....3). and cI (t) = Martingale Zt 0 c (u)dB (u): I (t) is a martingale. j =0 (tj ) B (tj +1 ) . j =0 (tj ) B (tj +1 ) .e. B(t` )) = (t` ) B (s) . B (tk )] We compute conditional expectations: 2` 1 3 `1 X X IE 4 (tj )(B (tj+1 ) .. Write I (t) = ` 1 X . there are partition points t ` and tk such that s 2 t` t`+1 ] and t 2 tk tk+1 ] (See Fig.. B (tj )] + (tk ) B (t) . j =0 j =0 IE (t`)(B (t`+1 ) . B (tk )] tk t tk+1 s t be given. B (t` )] k 1 X .

B (tj )) F (tj ) F (s) j =`+1 j =`+1 3 2 k 1 X 6 = IE 4 (tj ) (IE B (tj +1)jF (tj )] . so Zt 0 2 (u) du: I (t) = k X j =0 (tj ) B (tj +1 ){z B (tj )] . tj ) 2 (u) du = IE k X tZj+1 j =0 tj Zt 2 (u) du = IE 0 . B(tj )) F (s)7 5 {z } | j =`+1 =0 2 3 IE (tk )(B (t) . B(tk )) F (s) = IE 6 (tk ) |IE B(t)jF ({z )] . . B (tj ))2 F (tj ) IE 2(tj )(tj+1 . The Itˆ Integral o These first two terms add up to I (s). IEI 2(t) = = = IE 2(tj )Dj2] IE 2 (t j )IE (B (tj +1 ) .CHAPTER 14. 0k 12 X I 2(t) = @ (tj )Dj A j =0 k X 2 X 2 = (tj )Dj + 2 j =0 i<j k X j =0 k X j =0 k X j =0 (ti ) (tj )DiDj : Since the cross terms have expectation zero. 161 2k 1 3 k1 X X IE 4 (tj )(B (tj +1 ) . We show that the third and fourth terms are zero. B(tk )) F (s)7 tk 4 ( 5 } . . and different Dj are independent. =0 Theorem 7. } | Dj Each Dj has expectation 0. assume t = t k . B (tj )) F (s)5 = IE IE (tj )(B (tj +1 ) .46 (Itˆ Isometry) o IEI 2(t) = IE Proof: To simplify notation.

In the last section we have defined In (T ) = for every n. 14. 14.47 There is a sequence of elementary processes f n g1 such that n=1 nlim IE ! 1 ZT 0 j n(t) . Let IE be a process (not necessarily an elementary process) such that (t) is F (t)-measurable. 8t 2 0 T ].4 shows the main idea. R T 2(t) dt < 1: 0 Theorem 8. over 0 T ].4: Approximating a general process by an elementary process 4 . (t)j2 dt = 0: Proof: Fig. We now define ZT 0 n (t) dB (t) ZT 0 (t) dB (t) = nlim ! ZT 0 1 n (t) dB (t): .162 path of δ 4 path of δ 0=t0 t1 t2 t3 t4 = T Figure 14.8 Itˆ integral of a general integrand o Fix T > 0.

Linearity. square-integrable process. If I (t) = then Zt 0 (u) dB (u) J (t) = ( (u) Zt 0 (u) dB (u) I (t) J (t) = Zt 0 (u)) dB (u) and cI (t) = Zt 0 c (u)dB (u): Martingale. Continuity. IEI 2(t) = IE 0 2 (u) du.5 .CHAPTER 14. m (t)] dB (t) 2 n (t) . ˆ 14. m(t)j ]2 dt ZT ZT 2a2 + 2b2 :) 2IE j n (t) .1 () Consider the Itˆ integral o ZT 0 B(u) dB(u): We approximate the integrand as shown in Fig. (t)j2 dt + 2IE j m (t) . Suppose n and m are large positive integers. o Example 14. This guarantees that the sequence fI n (T )gn=1 has a limit.9 Properties of the (general) It o integral I (t) = Zt 0 (u) dB (u): Here is any adapted. I (t) is F (t)-measurable. Adaptedness. m (t)] dt !2 ((a + b)2 j n(t) . Rt Itˆ Isometry. For each t. I (t) is a continuous function of the upper limit of integration t. Im (T )) = IE (Itˆ Isometry:) = IE o ZT 0 T Z 0 0 n (t) . (t)j2 dt = IE 0 0 1 ZT which is small. I (t) is a martingale. Then var(In (T ) . (t)j + j (t) . The Itˆ Integral o 163 The only difficulty with this approach is that we need to make sure the above limit exists. 14.

X Bj2 . X B kT n k =0 B (k + 1)T .1 X B(u) dB(u) = nlim !1 k=0 Bk (Bk+1 . X Bk Bk+1 + 1 2 . . Bk ): . 1 2 2 k =0 Bk (Bk+1 .164 T/4 2T/4 3T/4 T Figure 14. n 1 j =0 n 1 k =0 . X 2 Bk+1 . X . ZT 0 B(u) dB(u) = nlim !1 n 1 . Bk ): 0 n 1 k =0 We compute 1 2 n 1 k =0 . X Bk Bk+1 + 1 2 n 1 k =0 n 1 k =0 . X 2 Bk n 1 k =0 n 1 . n 1 k =0 . X Bk Bk+1 = Bn . we denote so 4 Bk = B kT n ZT n. u < T: By definition. Bk )2 = 1 2 . 8 >B(0) = 0 > <B(T=n) n (u) = >: : : > (n. X 2 Bk . X 2 Bk n 1 k =0 1 2 = 2 Bn + 1 2 1 2 = 2 Bn + .1)T : B T if if if 0 u < T=n T=n u < 2T=n (n 1)T n . B kT n n : To simplify notation.5: Approximating the integrand B (u) with 4 . X (Bk+1 . over 0 T ].

4 (Reason for the 1 T term) If f is differentiable with f (0) = 0. 1 T: 2 2 The extra term 1 T comes from the nonzero quadratic variation of Brownian motion. 165 n 1 k =0 . 2 2 . for Brownian motion. 1 2 2 n 1 k =0 . then 2 ZT 0 f (u) df (u) = ZT 0 f (u)f (u) du 0 = 1 f 2 (u) 2 In contrast. X (Bk+1 .CHAPTER 14. we have T 0 = 1 f 2 (T ): 2 ZT 0 B (u)dB(u) = 1 B 2(T ) .10 Quadratic variation of an Itˆ integral o Theorem 10. Bk ) = 1 Bn . It has to be 2 there. X 2 Bk (Bk+1 . Bk )2 n 1 k =0 or equivalently n 1 . 1 T: 2 Remark 14. because Z IE T 0 B (u) dB (u) = 0 (Itˆ integral is a martingale) o but IE 1 B 2 (T ) = 1 T: 2 2 14. X B kT n k =0 B (k + 1)T . The Itˆ Integral o Therefore. X B (k + 1)T n k T 2 : Let n!1 and use the definition of quadratic variation to get ZT 0 1 B(u) dB(u) = 2 B 2 (T) . B kT n n 1 = 1 B 2 (T) .48 (Quadratic variation of Itˆ integral) Let o I (t) = Then Zt 0 (u) dB (u): 0 2(u) du: hI i(t) = Zt .

Proof: (For an elementary process ).. k=0 n 1 tZ +1 X k k=0 tk 2(t )(t k k+1 . hI i(tk ).e. B(sj )]2 k )(tk+1 . we can write the quadratic variation formula in differential form as follows: dI (t) dI (t) = 2(t) dt: Compare this with dB(t) dB(t) = dt: (t) = n 1 X . I (sj ) = so (tk ) dB (u) = (tk ) B (sj +1 ) .. B (sj )] hI i(tk+1) . j =0 I (sj+1) . To simplify notation. Let k=0 hI i(tk+1) . We have hI i(t) = Let us compute hI i(tk+1) . jj!! . the instantaneous absolute volatility of I is 2 (u). hI i(tk)] : tk = s0 Then = fs0 s1 : : : sm g be a partition s1 : : : sm = tk+1 : sZ+1 j sj I (sj+1 ) . tk ): hI i(t) = = n 1 X .e. (t)) in the Brownian motion.. Informally. This is the absolute volatility of the Brownian motion scaled by the size of the position (i.. The quadratic variation formula says that at each time u.166 This holds even if is not an elementary process.. I (sj )]2 m 1 X . Zt . tk ) 2 (u) du 0 2(u) du: jj 0 .. = 2 (tk ) jj jj!0 . .! It follows that j =0 2 (t B(sj+1 ) ... assume t = tn . i.. (tk ) for tk t tk+1 . Let = ft0 t1 : : : tn g be the partition for . hI i(tk) = m 1 X .

and in particular has nonzero quadratic variation. | (B (0)) = f {z } f (0) Zt 0 1 f (B(u)) dB (u) + 2 0 Zt 0 f (B (u)) du: 00 Remark 15. we obtain Itˆ ’s formula in integral form: o o f (B(t)) . Integral Forms) The mathematically meaningful form of Itˆ ’s foro mula is Itˆ ’s formula in integral form: o f (B(t)) . dt df (B(t)) = f (B(t)) dB (t) + 1 f (B(t)) |{z} : 2 0 00 dB (t) dB (t) This is Itˆ ’s formula in differential form. then the ordinary chain rule would give d f (B(t)) = f (B(t))B (t) dt 0 0 which could be written in differential notation as df (B (t)) = f (B(t))B (t) dt = f (B (t))dB (t) 0 0 0 However. f (B (0)) = Zt 0 f 0 1 (B (u)) dB (u) + 2 0 167 Zt f (B (u)) du: 00 .Chapter 15 Itˆ ’s Formula o 15. B (t) is not differentiable. namely. If B (t) were also differentiable. so the correct formula has an extra term.1 (Differential vs. Integrating this. where f (x) is a differentiable function.1 Itˆ ’s formula for one Brownian motion o We want a rule to “differentiate” expressions of the form f (B (t)).

Taylor’s formula to second order is exact because f is a quadratic function. xk )3 . This formula becomes mathematically respectable only after we integrate it. xk )2f (xk ): 2 In the general case. > 0 and let = ft0 t1 : : : tng be a partition of 0 T ]. The total error will have limit zero in the last step of the following argument. k=0 B(tk ) B(tk+1 ) . B(tk )]2 : . k=0 B (tk+1 ) . B(tk )] + 1 2 n 1 X . o Zt 0 f (B (u)) du 00 is a Riemann integral. the type used in freshman calculus. xk )f (xk ) + 1 (xk+1 . f (B (tk ))] B(tk+1 ) . Zt 0 f (B(u)) dB (u) 0 is an Itˆ integral. For paper and pencil computations. f (B(0)) 1 = 2 B 2 (T ) . Fix T In this case. 15. The first. 1 B 2 (0) 2 = = = n 1 X . Using Taylor’s formula. f (B(tk+1 )) .168 This is because we have solid definitions for both integrals appearing on the right-hand side. Taylor’s formula implies f (xk+1) . . k=0 n 1 X k=0 n 1 X . so that 2 Let xk f (x) = x f (x) = 1: 0 00 0 00 xk+1 be numbers. the above equation is only approximate. the more convenient form of Itˆ ’s rule is Itˆ ’s formula in differo o ential form: df (B (t)) = f (B (t)) dB(t) + 1 f (B (t)) dt: 2 There is an intuitive meaning but no solid definition for the terms df (B (t)) dB (t) and dt appearing 0 00 in this formula. and the error is of the order of (x k+1 . defined in the previous chapter.2 Derivation of Itˆ ’s formula o Consider f (x) = 1 x2 . we write: f (B(T )) . B (tk )] f (B (tk )) + 1 2 0 n 1 X . f (xk ) = (xk+1 . B(tk )]2 f (B(tk )) 00 k=0 B(tk+1 ) . The second.

3 Geometric Brownian motion Definition 15.CHAPTER 15.1 (Geometric Brownian Motion) Geometric Brownian motion is S (t) = S (0) exp B(t) + where Define and n > 0 are constant. Itˆ ’s Formula o We let jj 169 jj!0 to obtain f (B (T )) . f (B(0)) = = ZT ZT 0 0 B(u) dB(u) + 1 hB{zT ) 2 | i( } f 0 T ZT (B (u)) dB (u) + 1 f 2 0 | 00 (B (u)) du: {z } 1 This is Itˆ ’s formula in integral form for the special case o f (x) = 1 x2 : 2 15.2 2 t o n . o .1 2 2 f fx = f fxx = 2f: dS (t) = df (t B(t)) 1 = ft dt + fx dB + 2 fxx dBdB | {z } =( . Geometric Brownian motion in differential form is dt dS (t) = S (t)dt + S (t) dB (t) and Geometric Brownian motion in integral form is S (t) = S (0) + Zt 0 S (u) du + Zt 0 S (u) dB(u): . dt + f dB + 1 2 f dt 2 = S (t)dt + S (t) dB (t) 1 2 )f 2 Thus. f (t x) = S (0) exp x + 1 .1 2 2 t o so S (t) = f (t B (t)): Then ft = According to Itˆ ’s formula.

It has quadratic variation hGi(t) = Zt 0 Thus the quadratic variation of S is given by the quadratic variation of G. Stock is modelled by a geometric Brownian motion: dS (t) = S (t)dt + S (t)dB(t): . 15. T1 k=0 S (tk+1) . T1 T1 2 S 2(T ) ' 1 1 )]2 1 ZT2 2 S 2(u) du As T2 # T1 . we write dS (t) dS (t) = ( S (t)dt + S (t)dB(t))2 = 2S 2 (t) dt 15. This term has zero quadratic variation. The Itˆ integral o G(t) = Zt 0 is not differentiable.5 Volatility of Geometric Brownian motion Fix 0 T1 T2. the instantaneous relative volatility of S is 2 . holds (t) shares of stock. S (tk ' T2 . In differential notation.6 First derivation of the Black-Scholes formula Wealth of an investor. In other words. S (t) = S (0) + the Riemann integral Zt 0 S (u) du + Zt 0 S (u) dB(u) F (t) = is differentiable with F 0 (t) = Zt 0 S (u) du S (u) dB (u) 2 S 2(u) du: S (t).4 Quadratic variation of geometric Brownian motion In the integral form of Geometric Brownian motion. T2 . An investor begins with nonrandom initial wealth X 0 and at each time t. Let = volatility of S on T 1 T2] is ft0 : : : tng be a partition of T1 T2]. This is usually called simply the volatility of S . the above approximation becomes exact.170 15. The squared absolute sample n 1 X .

(t)S (t)] dt = (t) S (t)dt + S (t)dB (t)] + r X (t) . Let X (t) denote the wealth of the investor at time t. we obtain: vt + Svx + 1 2 S 2vxx = rX + ( . (t)S (t)] dt = rX (t)dt + (t)S (t) ( {z r) dt + (t)S (t) dB (t): | . } Risk premium Value of an option. Making these substitutions. the value of the option at each time t 2 0 T ] is v (t S (t)): The differential of this value is dv (t S (t)) = vt dt + vxdS + 1 vxxdS dS 2 = vt dt + vx S dt + S dB ] + 1 vxx 2 S 2 dt h i 2 1 2 S 2v = vt + Svx + 2 xx dt + SvxdB A hedging portfolio starts with some initial wealth X 0 and invests so that the wealth X (t) at each time tracks v (t S (t)). we should let v be the solution to the Black-Scholes partial differential equation 1 vt (t x) + rxvx (t x) + 2 2 x2vxx(t x) = rv (t x) satisfying the terminal condition If an investor starts with X 0 = v (0 S (0)) and uses the hedge (t) = vx (t X (t) = v(t S (t)) for all t. Consider an European option which pays g (S (T )) at time T . Itˆ ’s Formula o lending at interest rate r. 171 The investor finances his investing by borrowing or (t) can be random. and we are seeking to cause X to agree with v . Then dX (t) = (t)dS (t) + r X (t) . In other words. We saw above that dX (t) = rX + ( . and in particular. we obtain v (T x) = g (x): S (t)). but must be adapted.CHAPTER 15. we equate coefficients in their differentials. 1 vt + Svx + 2 2S 2vxx = rv + vx ( . r)S: 2 But we have set = vx . r)S ] dt + S dB: To ensure that X (t) = v (t S (t)) for all t. Let v (t x) denote the value of this option at time t if the stock price is S (t) = x. X (T ) = g (S (T )). then he will have . we obtain the -hedging rule: (t) = vx (t S (t)): Equating the dt coefficients. r)S (where v = v (t S (t)) and S = S (t)) which simplifies to vt + rSvx + 1 2 S 2vxx = rv: 2 In conclusion. Equating the dB coefficients.

172 15. We obtain We apply Itˆ ’s formula to compute dr o dr2(t) = df (r(t)) = f (r(t)) dr(t) + 1 f (r(t)) dr(t) dr(t) 2 0 00 r(t) = r(0) + a (b . 2acr2(t) dt + 2 r 2 (t) dB (t) The mean of r(t). then IEr(t) = b for every t. In integral form. 1): c r(0) . acIEr(t) which implies that d heact IEr(t)i = eact acIEr(t) + d IEr(t) = eact ab: dt dt eactIEr(t) . cIEr(u)) du: 0 Differentiation yields Zt d dt IEr(t) = a(b . If r(0) 6= b . The integral form of the CIR equation is r(t) = r(0) + a (b . where f (x) = x2 . cr(u)) du + Zt Z tq r(u) dB (u): = 2r(t) a(b . cr(u)) du + 0 Zt Z tq 0 r(u) dB (u): Taking expectations and remembering that the expectation of an Itˆ integral is zero. cIEr(t)) = ab . This is df (r(t)). then r(t) exhibits mean reversion: c c IEr(t) = b + e c act lim IEr(t) = b : t! 1 . cr(t)) dt + q r(t) dB (t) + a(b . cr(t)) dt + 3 q r(t) dB (t) 2 = 2abr(t) dt . r(0) = ab Integration yields We solve for IEr(t): Zt 0 . 2acr2(t) dt + 2 r 2 (t) dB (t) + 2 r(t) dt 3 = (2ab + 2 )r(t) dt . cr(t))dt + where a q r(t) dB(t) bc and r(0) are positive constants. eacu du = b (eact .7 Mean and variance of the Cox-Ingersoll-Ross process The Cox-Ingersoll-Ross model for interest rates is dr(t) = a(b . we obtain o IEr(t) = r(0) + a (b . this equation is 0 0 2(t). b : c c b If r(0) = c .

Itˆ ’s Formula o Variance of r(t). . . the random vector B (t) is F (t)-measurable. (IEr(t))2 2 2 b 2 = 2ac2 + r(0) . b act 2 c2 2ac c ac + c e 2 2 2 + r(0) . Associated with a d-dimensional Brownian motion. b e 2act + ac 2bc . 2acIEr2(t) dt which implies that d e2actIEr2(t) = e2act 2acIEr2(t) + d IEr2(t) dt dt 2act (2ab + 2)IEr(t): =e Using the formula already derived for IEr(t) and integrating the last equation. then the processes Bi (t) and Bj (t) are independent. . we have a filtration fF (t)g such that For each t. . the vector increments B(t1 ) . r(0) e 2act : c ac var r(t) = IEr2(t) .2 (d-dimensional Brownian Motion) A d-dimensional Brownian Motion is a process B (t) = (B1 (t) : : : Bd (t)) with the following properties: Each Bk (t) is a one-dimensional Brownian motion. . . The integral form of the equation derived earlier for dr2(t) is 173 r2 (t) = r2(0) + (2ab + 2 ) Taking expectations. For each t t1 : : : tn . B(tn 1 ) are independent of F (t). 2ac Zt 0 IEr2(u) du: d IEr2(t) = (2ab + 2)IEr(t) .CHAPTER 15. 2ac Zt 0 r2(u) du + 2 Zt 0 2 r 3 (u) dB(u): IEr2(t) = r2(0) + (2ab + 2) Differentiation yields Zt 0 IEr(u) du . r(0) e 2act : c IEr2(t) = ! .8 Multidimensional Brownian Motion Definition 15. 15. B(t) : : : B(tn) . If i 6= j . b ac e act + ac 2bc . we obtain Zt 0 r(u) du . after considerable algebra we obtain 2 2b b 2 + b2 + r(0) .

49 If i 6= j . k=0 Bi (tk+1) .174 15.9 Cross-variations of Brownian motions Because each component Bi is a one-dimensional Brownian motion. Bi (tk )] Bj (tk+1) . Bj (tk )] All the increments appearing in the sum of cross terms are independent of one another and have mean zero. For i 6= j . k=0 n 1 X . Bi(tk )]2 Bj (tk+1 ) . Bj (t` )] : Bi (tk+1 ) . tk )2 jj jj n 1 X . so C converges to the constant IEC . and each has expectation (tk+1 . It follows that var(C ) = As jj n 1 X . we have var(C )!0. Bj (tk )] : The increments appearing on the right-hand side of the above equation are all independent of one another and all have mean zero. var(C ) = IEC 2 = IE n 1 X . k=0 (tk+1 . Bi (tk )] Bj (tk+1 ) . Bi (tk )]2 and Bj (tk+1 ) . Therefore. tk ) = jj jj:T: = 0. First note that C2 = +2 n 1 X . Bi(tk+1 ) . k=0 (tk+1 . Bi (t` )] Bj (t`+1 ) . we have: Theorem 9. Therefore. define the sample cross variation Proof: Let = ft0 : : : tn g be a partition of of Bi and Bj on 0 T ] to be C = n 1 X . dBi(t) dBj (t) = 0 0 T ]. Bj (tk )]2 are independent of one another. jj!0. Bj (tk )]2 : But Bi (tk+1 ) . tk ). Bi (tk ) Bj (tk+1) . we have the informal equation dBi(t) dBi (t) = dt: However. IEC = 0: We compute var(C ). k=0 Bi (tk+1 ) . Bj (tk ) 2 2 `<k Bi (t`+1) .

The integrands (u) (u) and ij (u) can be any o adapted processes. we write dX = dt + 11 dB1 + 12 dB2 dY = dt + 21 dB1 + 22 dB2 : Given these two semimartingales X and Y . are called semimartingales. 0 11 fx + 21 fy ] dB1 + Zt 0 2 1 2 11 21 + 12 22)fxy + 2 ( 21 + 22)fyy ] du 12 fx + 22 fy ] dB2 . and Bi = Bi (u). consisting of a nonrandom initial condition. plus one or more Itˆ integrals. Let X and Y be processes of the form X (t) = X (0) + Y (t) = Y (0) + Zt 0 Z0t (u) du + (u) du + Zt Z0t 0 11(u) dB1 (u) + 21 (u) dB1 (u) + Zt Z0t 0 12 (u) dB2(u) 22 (u) dB2 (u): Such processes. plus a Riemann integral. with X and Y as decribed earlier and with all the variables filled in. ij = ij (u). we write the Itˆ formula for two processes driven by a o two-dimensional Brownian motion. for i j 2 f1 2g. Then we have the corresponding Itˆ formula: o df (t x y) = ft dt + fx dX + fy dY + 1 fxx dX dX + 2fxy dX dY + fyy dY dY ] : 2 In integral form. and let X (t) and Y (t) be semimartingales.10 Multi-dimensional Itˆ formula o To keep the notation as simple as possible. this equation is f (t X (t) Y (t)) .CHAPTER 15. Itˆ ’s Formula o 175 15. In differential notation. the quadratic and cross variations are: dX dX = ( dt + 11 dB1 + 12 dB2)2 2 2 = 11 dB1{z 1 +2 11 12 dB1{z 2 + 12 dB2{z 2 | dB } | dB} | dB} =( dY dY = ( =( dX dY = ( =( 0 dt 2 + 2 )2 dt 11 12 dt + 21 dB1 + 22 dB2 )2 2 2 2 21 + 22 ) dt dt + 11 dB1 + 12 dB2)( 11 21 + 12 22 ) dt dt dt + 21 dB1 + 22 dB2 ) Let f (t x y ) be a function of three variables. The formula generalizes to any number of processes driven by a Brownian motion of any number (not necessarily the same number) of dimensions. f (0 X (0) Y (0)) Zt 2 2 = ft + fx + fy + 1 ( 11 + 12)fxx + ( 2 + where f 0Z t = f (u X (u) Y (u). The adaptedness of the integrands guarantees that X and Y are also adapted.

176 .

yj t0 x y.1 Stochastic Differential Equations Consider the stochastic differential equation: dX (t) = a(t X (t)) dt + (t X (t)) dB (t): Here a(t x) and (t x) are given functions. then you can evaluate X (t). Example 16. a(t y)j Ljx .1 (Drifted Brownian motion) Let a be a constant and = 1.. (t y)j Ljx . If you know the path of the Brownian motion up to time t. A solution to (SDE) with the initial condition (t 0 x) is a process fX (t)gt X (t0) = x X (t) = X (t0) + a(s X (s)) ds + t0 Zt Zt t0 (s X (s)) dB (s) t t0 The solution process fX (t)g t t0 will be adapted to the filtration fF (t)g t 0 generated by the Brownian motion. usually assumed to be continuous in (t chitz continuous in x.e. Let (t0 x) be given. yj for all t satisfying j (t x) .Chapter 16 Markov processes and the Kolmogorov equations 16. so dX(t) = a dt + dB(t): If (t0 x) is given and we start with the initial condition X(t0 ) = x 177 . there is a constant L such that (SDE) x) and Lips- ja(t x) .i.

t. t. and compute the expected value of h(X (t1)). B(t0 )) To compute the differential w. Denote by IE t0 xh(X (t1)) = x. Consider dX(t) = rX(t) dt + X(t) dB(t): Given the initial condition X(t0 ) = x the solution is X(t) = x exp (B(t) . treat t0 and B(t0 ) as constants: dX(t) = (r . and start with initial the expectation of h(X (t1)). Now let 2 IR be given. and based on this information. you want to estimate h(X (t 1)). t0) + (B(t) . the only relevant information is the value of X (t0). t0) : 2 Again.r. 1 2)X(t) dt + X(t) dB(t) + 1 2 X(t) dt 2 2 = rX(t) dt + X(t) dB(t): 16.2 (Geometric Brownian motion) Let r and be constants.t. if you observe the path of the driving Brownian motion from time 0 to time t 0 . B(t0 )) + (r . to compute the differential w.178 then X(t) = x + a(t . . You imagine starting the (SDE ) at time t0 at value X (t0). given that X (t0) condition X (0) = : We have the Markov property IE 0 h(X (t1)) F (t0) = IE t0 X (t0)h(X (t1)): In other words. treat t0 and B(t0 ) as constants: dX(t) = a dt + dB(t): t t0: Example 16.r. 1 2 )(t .2 Markov Property Let 0 t0 < t1 be given and let h(y ) be a function. t.

2 2 )(t1 .3 Transition density Denote by p(t0 t1 x y) the density (in the y variable) of X (t 1). t0) : 2 dy = y db or equivalently. X(t 0 ) = x. t0 ) 2 or equivalently. 1 2)(t1 . 2(t b. IE t0 xh(X (t1)) = The Markov property says that for 0 Z IR h(y )p(t0 t1 x y ) dy: . Markov processes and the Kolmogorov equations 179 16. t0 ))) : 2(t1 . is Geometric Brownian motion: 1 X(t1 ) = x exp (B(t1 ) . t0 ) and variance 2 p(t0 t1 x y) = p 1 exp . t0 ). This is always the case when a(t x) and (t x) don’t depend on t. the random variable X(t1 ) is normal with mean x + a(t1 . B(t0 ) has density 2 IP fB(t1 ) . B(t0 )) + (r .e. (y . 1 2)(t1 . t0) : The random variable B(t1 ) . The derivative is db = dy : y . (r . t0 . In other words.CHAPTER 16. (x + a(t1 . t0 t1 and for every IE 0 h(X (t1)) F (t0 ) = Z IR h(y )p(t0 t1 X (t0) y ) dy: Example 16. conditioned on X (t 0) = x. Example 16.4 (Geometric Brownian motion) Recall that the solution to the SDE dX(t) = rX(t) dt + X(t) dB(t) with initial condition X(t 0 ) = x. i..3 (Drifted Brownian motion) Consider the SDE dX(t) = a dt + dB(t): Conditioned on (t1 . t0) 1 0 and we are making the change of variable y = x exp b + (r . t0 ) 2 (t1 . t ) db 2 (t1 . t0) Note that p depends on t0 and t1 only through their difference t 1 . B(t0 ) 2 dbg = p 1 exp . h y i b = 1 log x .

2 x dx: 2 .t log X(t) + r(T .t 16. K)+ 1 = X(t)N p 1 log X(t) + r(T . Then the Kolmogorov Backward Equation is: @ @ @2 1 . (r .r(T .r(T . We then write p( x y ) rather than p(t0 t1 x y ). and (KBE ) becomes The variables t0 and x in (KBE ) are called the backward variables.r(T . 2 x dx = p1 e. K) = tx 2 h y i2 1 log x .4 The Kolmogorov Backward Equation Consider and let p(t0 t1 x y ) be the transition density. t0 . Z Z1 12 1 2 e. .t) (X(T) .t) K N p 1 2 K T . e.t Z1 where N( ) = p1 2 Therefore. t) T . t ) = p 1 y 2 (t1 . t) . K)+ p(t T x y) dy 0 h x i r (T . 2(t .1 IE 0 e. t) K T . @ p( x y ) = a(x) @ p( x y) + 1 2(x) @ 2 p( x y ): 2 @ @x @x2 (KBE’) . p(t0 t1 x y) dy = IP fX(t1 ) 2 dyg 1 exp . t) . 1 2 (T .180 Therefore. one can compute the expected payoff from a European call: + (y . K)+ F (t) = e. t) + 2 2(T . t0 ) dy: Using the transition density and a fair amount of calculus. 1 2 (T . @t p(t0 t1 x y) = a(t0 x) @x p(t0 t1 x y) + 2 2(t0 x) @x2 p(t0 t1 x y): 0 (KBE) dX (t) = a(t X (t)) dt + (t X (t)) dB (t) In the case that a and are functions of x alone. t) + 2 2(T . t0) 1 0 IE (X(T ) .t) IE t X (t) (X(T) .t h x i log K + r(T . KN p 1 2 T .t) 1 =e xN p 1 log K + r(T . t) . t) . 2 2 )(t1 . p(t0 t1 x y ) depends on t0 and t1 only through their difference = t1 .

and define (5. x . (y . a )2 . (y . a p: x @x @2 p = p = @ y .5 (Drifted Brownian motion) 181 dX(t) = a dt + dB(t) 2 p( x y) = p 1 exp . @ 2 2 2 1 + a(y . 1 p + (y . 2 2 ) y 2 It is true but very tedious to verify that p satisfies the KBE 1 p = rxpx + 2 2x2 pxx: i2 : 16. (y . a )2 @ @ 2 2 @ (y . Example 16. a ) + (y . x . x . x . x . (x + a )) : 2 2 @ p = p = @ p 1 exp .CHAPTER 16.1) v (t x) = IE t xh(X (T )) .2 22 @ p = p = y . a )2 p 1 exp . a ) p 2 2 =p : This is the Kolmogorov backward equation. 2 1 2 log x . Markov processes and the Kolmogorov equations Example 16. x . a ) p: 2 Therefore. a ) p: = . x . x . a ) . x . 21 + (y . a p xx x @x2 @x 2 = . a p + y . (r . 2 1 apx + 2 pxx = a(y .6 (Geometric Brownian motion) dX(t) = rX(t) dt + X(t) dB(t): h y 1 1 p( x y) = p exp .5 Connection between stochastic calculus and KBE Consider dX (t) = a(X (t)) dt + (X (t)) dB(t): Let h(y ) be a function. x . x .

t x y ) dy = 0 2 Let (0 IE 0 . the process v (t X (t)) satisfies the martingale property: 0 s t T: IE v (t X (t)) F (s) = v (s X (s)) Proof: According to the Markov property.1).p (T . h(y ) p (T . We simplify notation by writing IE rather than Theorem 5. t x y ) + 1 2(x)pxx(T . ) be an initial condition for the SDE (5.50 Starting at X (0) = . t x y ) dy vx(t x) = h(y) px(T . the Kolmogorov backward equation implies Z Z Z Z vt(t x) +Za(x)vx(t x) + 1 2(x)vxx(t x) = 2 h i h(y ) . Then v(t x) = h(y) p(T . t x y ) dy vt(t x) = . t x y) + a(x)px(T .182 where 0 t T . IE h(X (T )) F (t) = IE t X (t)h(X (T )) = v (t X (t)) so IE v (t X (t))jF(s)] = IE IE h(X (T )) F (t) F (s) = IE h(X (T )) F (s) = IE s X (s)h(X (T )) = v (s X (s)): (Markov property) Itˆ ’s formula implies o dv(t X (t)) = vt dt + vxdX + 1 vxx dX dX 2 1 = vt dt + avx dt + vx dB + 2 2 vxx dt: . t x y ) dy: Therefore. t x y ) dy vxx(t x) = h(y) pxx(T .

we have come to the same conclusion. hence h 1 vt + avx + 2 2 vxx du must be zero i vt + avx + 1 2vxx = 0: 2 Thus by two different arguments. o Theorem 5. Markov processes and the Kolmogorov equations In integral form. In fact. as we show in the next section.CHAPTER 16. one can use Itˆ ’s formula to prove the Feynman-Kac Theorem. 1 2)(u .6 Black-Scholes Consider the SDE With initial condition the solution is dS (t) = rS (t) dt + S (t) dB (t): S (t) = x S (u) = x exp (B (u) . 16. and use o the Feynman-Kac Theorem to derive the Kolmogorov backward equation. so the integral 0 for all t. B(t)) + (r . This implies that the integrand is zero.51 (Feynman-Kac) Define v(t x) = IE t xh(X (T )) where Then 0 t T dX (t) = a(X (t)) dt + (X (t)) dB(t): vt(t x) + a(x)vx(t x) + 1 2 (x)vxx(t x) = 0 2 (FK) and v (T x) = h(x): The Black-Scholes equation is a special case of this theorem. one based on the Kolmogorov backward equation. we have 183 v (t X (t)) = v (0 X (0)) Z th i + vt (u X (u)) + a(X (u))vx(u X (u)) + 1 2 (X (u))vxx(u X (u)) du 2 + Z0t 0 (X (u))vx(u X (u)) dB (u): Rt We know that v (t X (t)) is a martingale.1 (Derivation of KBE) We plunked down the Kolmogorov backward equation without any justification. Remark 16. t) 2 n o u t: . and the other based on Itˆ ’s formula.

1 2 )(T . B (t)) + (r . t) |{z} F n t T . B (t)) + (r . IE h(X Y ) G = (X ) where (x) = IEh(x Y ): With geometric Brownian motion. B (t)) + (r . 1 2)(T .184 Define v (t x) = IE t xh(S (T )) n o = IEh x exp (B (T ) . for 0 S (t) = S (0) exp B(t) + (r . t) : 2 Now IEh(xY ) = v(t x): The independence lemma implies IE h(S (T )) F (t) = IE h(XY )jF (t)] = v (t X ) = v (t S (t)): . then where h is a function to be specified later. 1 2)t 2 n o S (T ) = S (0) exp B(T ) + (r . 2 2 )(T . X is G -measurable. 1 2)T 2 n o 1 = S (t) exp (B (T ) . we have o (t)-measurable | independent of F (t) {z } We thus have S (T ) = XY where X = S (t) o n Y = exp (B(T ) . and Y is independent of G . t) 2 Recall the Independence Lemma: If G is a -field.

we have the terminal condition v(T x) = h(x) x 0: Furthermore. . we shall eventually see that the value at time t of a contingent claim paying h(S (T )) is v(t 0) = h(0) . is a martingale: For 0 s t T . then also S (T ) = 0. the tower property implies that v (t S (t)) 0 t T . .rer(T t)u(t x) + er(T t)ut (t x) x) = er(T t)ux(t x) x) = er(T t)uxx(t x): . . Markov processes and the Kolmogorov equations We have shown that 185 v(t S (t)) = IE h(S (T )) F (t) 0 t T: Note that the random variable h(S (T )) whose conditional expectation is being computed does not depend on t. IE v(t S (t)) F (s) = IE IE h(S (T )) F (t) F (s) = IE h(S (T )) F (s) = v (s S (s)): This is a special case of Theorem 5. . Therefore. the sum of the dt terms in dv (t S (t)) must be 0. Because of this. 0 t < T x 0: Along with the above partial differential equation.51 (Feynman-Kac). Because formula. .51. . . dv (t S (t)) = vt(t S (t)) dt + rS (t)vx(t S (t)) + 1 2 S 2(t)vxx(t S (t)) dt 2 + S (t)vx(t S (t)) dB (t): By Itˆ ’s o h i This leads us to the equation vt(t x) + rxvx(t x) + 1 2x2vxx(t x) = 0 2 This is a special case of Theorem 5. if S (t) = 0 for some t 2 0 T ]. v(t v t (t vx(t vxx(t x) = er(T t)u(t x) x) = . r(T t) IE t xh(S (T )) r(T t) v (t x) . v(t S (t)) is a martingale. 0 t T: u(t x) = e =e at time t if S (t) = x. condition This gives us the boundary Finally.CHAPTER 16.

The Feynman-Kac Theorem now implies that . t) . t) ( 2 y 1 log x . y )+ . t) y 2 (T . (r .6. 2(T . t) 2 T . t) 2) for geometric Brownian motion (See Example 16. we have the “stochastic representation” u(t x) = e =e In the case of a call. . 1 p(t T x y ) = p 1 exp . K )+ x 0 0 t < T x > 0: .t Even if h(y ) is some other function (e.. a put). and . e r(T t)K N p 1 2 .7 Black-Scholes with price-dependent volatility x log K + r(T . h(y ) = (K .rv(t x) + vt(t x) + rxvx(t x) + 1 2(x)vxx(t x) = 0 2 v also satisfies the terminal condition v(T x) = (x . r(T t) IE t xh(S (T )) . 1 2 (T .ru(t x) + ut(t x) + rxux(t x) + 1 2x2uxx(t x) = 0 2 In terms of the transition density 0 t < T x 0: (BS) Compare this with the earlier derivation of the Black-Scholes PDE in Section 15. dS (t) = rS (t) dt + (S (t)) dB(t) v(t x) = e r(T t) IE t x(S (T ) .t) appearing in every term. . K )+ : . K )+ u(t x) = x N p1 satisfies the Black-Scholes PDE (BS) derived above. . t) . we obtain the Black-Scholes partial differential equation: . r(T t) Z (SR) 1 0 h(y )p(t T x y) dy: h(y ) = (y . T .4).g. . u(t x) is still given by and 16.t x log K + r(T . 2 2)(T .186 Plugging these formulas into the partial differential equation for v and cancelling the e r(T . t) + 1 2 (T .

CHAPTER 16. Markov processes and the Kolmogorov equations
and the boundary condition

187

v(t 0) = 0

0 t T:

An example of such a process is the following from J.C. Cox, Notes on options pricing I: Constant elasticity of variance diffusions, Working Paper, Stanford University, 1975:

dS (t) = rS (t) dt + S (t) dB (t)
where 0 < 1. The “volatility” sponding Black-Scholes equation is

S

;

1(t) decreases with increasing stock price. The corre-

;rv + vt + rxvx + 1 2x2 vxx = 0 2

0 t<T x>0 v (t 0) = 0 0 t T v (T x) = (x ; K )+ x 0:

188

Chapter 17

Girsanov’s theorem and the risk-neutral measure
(Please see Oksendal, 4th ed., pp 145–151.) Theorem 0.52 (Girsanov, One-dimensional) Let B (t) 0 t T , be a Brownian motion on a probability space ( F P). Let F (t) 0 t T , be the accompanying filtration, and let (t) 0 t T , be a process adapted to this filtration. For 0 t T , define

e B (t) =

Zt
0

(u) du + B (t)

Z (t) = exp ;

Zt
0

(u) dB (u) ;

Zt 2 1 2 0 (u) du 8A 2 F :

and define a new probability measure by

f IP (A) = f e Under I , the process B (t) P

Z
A

Z (T ) dIP

0 t T , is a Brownian motion.

Caveat: This theorem requires a technical condition on the size of . If

IE exp
everything is OK. We make the following remarks:

( ZT
1 2 0

2 (u) du

)

<1

Z (t) is a matingale. In fact, 1 dZ (t) = ; (t)Z (t) dB (t) + 2 2 (t)Z (t) dB(t) dB(t) ; 1 2(t)Z (t) dt 2 = ; (t)Z (t) dB (t):
189

190

f IP is a probability measure.

Since Z (0) = 1, we have IEZ (t) = 1 for every t

f IP ( ) =

Z

0. In particular

Z (T ) dIP = IEZ (T ) = 1

f IE in terms of IE .

f so I is a probability measure. P

f IEZ = IE Z (T )X ] : To see this, consider first the case X = 1 A , where A 2 F . We have Z Z f f IEX = IP (A) = Z (T ) dIP = Z (T )1A dIP = IE Z (T )X ] :
A
Now use Williams’ “standard machine”. The intuition behind the formula

f f E P Let I denote expectation under I . If X is a random variable, then

f IP and IP .

f IP (A) =

Z
A

Z (T ) dIP

8A 2 F f IP . Thus,

is that we want to have

e Distribution of B (T ). If

f but since IP (! ) = 0 and I (!) = 0, this doesn’t really tell us anything useful about P we consider subsets of , rather than individual elements of .
is constant, then

f IP (!) = Z (T !)IP (!) n

Z (T ) = exp ; B (T ) ; 1 2 T 2 e (T ) = T + B(T ): B e Under IP , B (T ) is normal with mean 0 and variance T , so B (T ) is normal with mean T and variance T : ( ) b 1 exp ; (~ ; T )2 d~: e IP (B (T ) 2 d~) = p b b 2T 2 T e f e P Removal of Drift from B (T ). The change of measure from IP to I removes the drift from B (T ).
To see this, we compute

o

Z = p1 2 T 1 Z =p 2 T Z (y = T + b) = p 1 2 T = 0:

fe IEB (T ) = IE Z (T )( T + B(T ))] h n o i = IE exp ; B (T ) ; 1 2 T ( T + B (T )) 2
1 ;1 1

( T + b) expf; b ;

;1 1

( T + b) exp ; (b + TT ) 2
2 y exp ; y2 dy

(

1 2 2 Tg

( 2) b exp ; 2T db 2)
db
= T + b)

(

)

;1

(Substitute y

CHAPTER 17. Girsanov’s theorem and the risk-neutral measure

191

fe E We can also see that I B (T ) = 0 by arguing directly from the density formula
2 ~ IP B (t) 2 db = p 1 exp ; (b ; TT ) d~: b 2 2 T
Because

ne

o ~

(

)

Z (T ) = expf; B(T ) ; 1 2T g 2 e (T ) ; T ) ; 1 2 T g = expf; (B 2 e (T ) + 1 2 T g = expf; B 2
we have

f e e IP B(T ) 2 d~ = IP B (T ) 2 d~ exp ; ~ + 1 2 T b b b 2 ( ~ ) 1 exp ; (b ; T )2 ; ~ + 1 2 T d~: b 2 =p b 2T 2 T ( ) b 1 exp ; ~2 d~: =p 2T b 2 T f e Under I , B (T ) is normal with mean zero and variance P mean T and variance T .
T.
Under IP ,

n

o

n

o

n

o

e B (T ) is normal with

Means change, variances don’t. When we use the Girsanov Theorem to change the probability measure, means change but variances do not. Martingales may be destroyed or created. Volatilities, quadratic variations and cross variations are unaffected. Check:

e e dB dB = ( (t) dt + dB(t))2 = dB:dB = dt: f 17.1 Conditional expectations under I P
Lemma 1.53 Let 0

t T . If X is F (t)-measurable, then

f IEX = IE X:Z (t)]:

Proof:

f IEX = IE X:Z (T )] = IE IE X:Z (T )jF (t)] ] = IE X IE Z (T )jF (t)] ] = IE X:Z (t)]
because Z (t)

0 t T , is a martingale under IP .

192 Lemma 1.54 (Baye’s Rule) If X is F (t)-measurable and 0

1 f IE X jF (s)] = Z (s) IE XZ (t)jF (s)]:

s t T , then
(1.1)

Proof: It is clear that Z 1s) IE XZ (t)jF (s)] is ( property. For A 2 F (s), we have

F (s)-measurable.

We check the partial averaging

Z

1 IE XZ (t)jF (s)] dIP = IE 1 1 IE XZ (t)jF (s)] f f A Z (s) A Z (s) (Lemma 1.53) = IE 1A IE XZ (t)jF (s)]] = IE IE 1A XZ (t)jF (s)]] (Taking in what is known) = IE 1A XZ (t)] f = IE 1A X ] (Lemma 1.53 again) Z f = X dIP :
A

Although we have proved Lemmas 1.53 and 1.54, we have not proved Girsanov’s Theorem. We will not prove it completely, but here is the beginning of the proof. Lemma 1.55 Using the notation of Girsanov’s Theorem, we have the martingale property

fe e IE B (t)jF (s)] = B (s)

0 s t T:

e Proof: We first check that B (t)Z (t) is a martingale under IP . Recall e dB (t) = (t) dt + dB (t) dZ (t) = ; (t)Z (t) dB(t):

Therefore,

e e e e d(BZ ) = B dZ + Z dB + dB dZ e = ;B Z dB + Z dt + Z dB ; Z dt e = (;B Z + Z ) dB:
s t T,

Next we use Bayes’ Rule. For 0

fe e IE B (t)jF (s)] = Z 1s) IE B (t)Z (t)jF (s)] ( e = Z 1s) B (s)Z (s) ( e = B (s):

CHAPTER 17. Girsanov’s theorem and the risk-neutral measure

193

Definition 17.1 (Equivalent measures) Two measures on the same probability space which have the same measure-zero sets are said to be equivalent. The probability measures defined by

IP

and

f IP of the Girsanov Theorem are equivalent.
A 2 F:

Recall that

f IP is

Z f(A) = Z (T ) dIP IP Z

If IP (A) = 0, then A Z (T ) f of I to obtain P

R

dIP = 0: Because Z (T ) > 0 for every !, we can invert the definition IP (A) = dIP A Z (T )
1

f

A 2 F:

f P If I (A) = 0, then A Z (1T )

R

dIP = 0:

17.2 Risk-neutral measure
As usual we are given the Brownian motion: B (t) 0 t T , with filtration F (t) defined on a probability space ( F P). We can then define the following. Stock price:

0

t

T,

dS (t) = (t)S (t) dt + (t)S (t) dB (t):
The processes (t) and (t) are adapted to the filtration. The stock price model is completely general, subject only to the condition that the paths of the process are continuous.

r(t) 0 t T . The process r(t) is adapted. Wealth of an agent, starting with X (0) = x. We can write the wealth process differential in
Interest rate: several ways:

dX (t) =

= r(t)X (t) dt + (t) dS (t) ; rS (t) dt] = r(t)X (t) dt + (t) ( (t) {z r(t)) S (t) dt + (t) (t)S (t) dB (t) | ; }

Capital gains from Stock

dS ) | (t){z (t} + r(t) X (t) ;{z (t)S (t)] dt | }
Interest earnings

2 3 6 (t) ; r(t) 7 6 7 6 = r(t)X (t) dt + (t) (t)S (t) 6 dt + dB (t)7 7 ( 4 | {zt) } 5
Risk premium Market price of risk=

(t)

Define (t) d S(t) = 1t) . r(t))S (t) dt + (t)S (t) dB (t)] ( e B (t) = Then Zt 0 (u) du + B (t): (t) e d S(t) = 1t) (t)S (t) dB (t) ( e d X((tt)) = ((tt)) (t)S (t) dB (t): f (t) Under I .2 (Risk-neutral measure) A risk-neutral measure (sometimes called a martingale measure) is any probability measure.194 Discounted processes: d e d e . S (t) and X((tt)) are martingales. P Definition 17. R t r(u) du 0 R t r(u) du 0 S (t) = e .r(t)S (t) dt + dS (t)] . The discounted formulas are = 1t) (t)S (t) (t) dt + dB (t)] ( (t) d X((tt)) = (t) d S(t) () = (tt) (t)S (t) (t) dt + dB (t)]: Changing the measure. equivalent to the market measure IP . .r(t)X (t) dt + dX (t)] S (t) : 0 = (t)d e Notation: R t r(u) du Rt (t) = e 0 r(u) du d (t) = r(t) (t) dt 1 = e R0t r(u) du (t) d 1t) = . R t r(u) du 0 . which makes all discounted asset prices martingales. r(tt) dt: ( () .r(t)S (t) dt + dS (t)] ( = 1t) ( (t) . R t r(u) du 0 . X (t) = e . .

e. Girsanov’s theorem and the risk-neutral measure For the market model considered here. then (t) 0 t T . (t) 6= Risk-neutral valuation.1 V = (S(T) . Note that because 0. Consider a contingent claim paying an F (T )-measurable random variable V at time T . a process satisfies X (T ) = V . K V= T Asian call 0 V = 0maxT S(t) Look back t If there is a hedging portfolio. whose corresponding wealth process f V X (0) = IE (T ) : f P This is because X((tt)) is a martingale under I .. 195 f IP (A) = where Z A 0 Z (T ) dIP (u) dB (u) . i. K)+ European call + V = (K . Zt Zt is the unique risk-neutral measure. so (0) f T f V X (0) = X(0) = IE X((T )) = IE (T ) : . 1 2 A2F Z (t) = exp . (t) = 2 (u) du 0 (t) r(t) we must assume that (t) . Example 17.CHAPTER 17. S(T)) European put !+ ZT 1 S(u) du .

196 .

2 A hedging application Homework Problem 4. Let X (t) 0 t relative to this filtration. Now we see that if X (t) is a martingale adapted to the filtration generated by the Brownian motion B (t). In the context of Girsanov’s Theorem. Then there is an adapted process (t) 0 F P). then dX (t) = (t) dB(t) for some (t). the paths of X are continuous. 4th ed. 18.56 Let B (t) 0 t T be a Brownian motion on ( the filtration generated by this Brownian motion.1 We already know that if X (t) is a process satisfying dX (t) = (t) dB (t) then X (t) is a martingale. p.11. Y (t) = Y (0) + Zt 0 e (u) dB (u) 197 0 t T: . Theorem 1. be T . be a martingale (under IP ) t T .50. i.1 Martingale Representation Theorem See Oksendal. Remark 18.e. Let F (t) 0 t T . such that 0 t T is f IP -martingale. the Brownian motion is the only source of randomness in X . suppse that F (t) the filtration generated by the Brownian motion B (under IP ).5. Suppose that Y is a Then there is an adapted process (t) 0 t T . such that X (t) = X (0) + Zt 0 (u) dB (u) 0 t T: In particular. Theorem 4..Chapter 18 Martingale Representation Theorem 18.

A 0 Zt (u) dB (u) . 1 2 Zt 0 2 (u) du 8A 2 F : Then Let (t) (t) e d S(t) = S(t) (t) dB (t): (t) 0 t T be a portfolio process. so that f P Define the I -martingale X (T ) = V: f V Y (t) = IE (T ) F (t) Zt 0 0 t T: (t) 0 t T . X (t) = X (0) + Z t (u) (u) S (u) dB(u) e (t) (u) 0 0 t T: Let V be an F (T )-measurable random variable..198 dS (t) = (t)S (t) dt + (t)S (t) dB (t) Zt (t) = exp r(u) du 0 (t) = (t) . representing the payoff of a contingent claim at time T . The corresponding wealth process X (t) satisfies (t) e d X((tt)) = (t) (t) S(t) dB(t) i.e.5. there is an adapted process f Set X (0) = Y (0) = I E h Y (t) = Y (0) + V (T ) and choose i e (u) dB (u) (u) so that (u) (u) S (u) = (u): (u) . such that 0 t T: According to Homework Problem 4. r(t) Z t (t) e B (t) = (u) du + B(t) 0 Z f(A) = Z (T ) dIP IP Z (t) = exp . We want to choose X (0) and (t) 0 t T .

CHAPTER 18. we have X (t) = Y (t) = IE V F (t) f (t) (T ) 0 t T: In particular. although it does not tell us how to compute it. 2 0 0 Zt jj (u)jj2 du . Zt 0 (u) dB (u) . perhaps larger than the one generated by B . Zt 0 (r(u) + 1 2 (u)) du 2 18. B(t) = (B1 (t) : : : Bd(t)) 0 t T . j (u) du + Bj (t) Zt j = 1 ::: d 1 (u): dB (u) . a d- e Bj (t) = Zt 0 Z f(A) = Z (T ) dIP: IP A Z (t) = exp . the accompanying filtration. It also justifies the risk-neutral pricing formula f X (t) = (t)IE = Z(t) IE (t) 1 IE = (t) where (T ) F (t) Z (T ) V F (t) (T ) (T )V F (t) V 0 t T (t) (t) = Z(t) = exp . X (T ) = IE V F (T ) = V f (T ) (T ) (T ) so X (T ) = V: The Martingale Representation Theorem guarantees the existence of a hedging portfolio. Martingale Representation Theorem With this choice of 199 (u) 0 u T . F P).57 (d-dimensional Girsanov) dimensional Brownian motion on ( F (t) 0 t T t T define For 0 (t) = ( 1(t) : : : d (t)) 0 t T . d-dimensional adapted process.3 d-dimensional Girsanov Theorem Theorem 3.

the process e e e B (t) = (B1(t) : : : Bd (t)) is a d-dimensional Brownian motion. then there is a d-dimensional adpated process (t) = ( 1(t) : : : d(t)) such that Z Y (t) = Y (0) + t 0 e (u): dB (u) 0 t T: 18. be a d-dimensional Brownian motion on some ( F P). is a martingale under I relative to F (t) 0 t T . be the filtration generated by B . Zt 0 (u): dB (u) 0 t T: Corollary 4.59 If we have a d-dimensional adapted process (t) = ( 1 (t) : : : d (t)) then we can e f f P P define B Z and I as in Girsanov’s Theorem. If Y (t) 0 t T .4 d-dimensional Martingale Representation Theorem Theorem 4. and let F (t) 0 t T . 0 t T 18. B(t) = (B1 (t) : : : Bd (t)) 0 t T a d-dimensional Brownian motion t T .5 Multi-dimensional market model following: B (t) = (B1(t) : : : Bd(t)) 0 t T . such that X (t) = X (0) + F (t) 0 t T P). Then we can define the Let Stocks dSi (t) = i (t)Si(t) dt + Si (t) Accumulation factor d X j =1 ij (t) dBj (t) i = 1 ::: m (t) = exp Here. under I . is a martingale (under IP ) relative to F (t) 0 If X (t) 0 d-dimensional adpated process (t) = ( 1(t) : : : d (t)).200 f P Then. then there is a the filtration generated by the Brownian motion B . .58 on ( F t T . i (t) Zt 0 r(u) du : ij (t) and r(t) are adpated processes.

then there is no risk-neutral probability measure and the market admits arbitrage. pp 313-316): The risk-neutral measure is unique if and only if every contingent claim can be hedged. every discounted stock price is a martingale. then it admits no arbitrage. Using (t) in the d-dimensional Girsanov Theorem. r(t) i = 1 : : : m: ::: Market price of risk. the market is said to be incomplete. Stochastic Proc. we need to choose 1 (t) ::: d (t).CHAPTER 18. and Applications 11 (1981). If (MPR) has multiple solutions. } ij (t) dBj (t) | j =1 d ? Si (t) X (t) (t) + dB (t)] = (t) ij | j {z j } j =1 e dBj (t) Risk Premium (5. Part II. Under I . Martingales and Stochastic integrals in the theory of continuous trading. Finally. f Consequently. so that (MPR) d X j =1 ij (t) j (t) = i (t) . Theorem 5. pp 215-260. the market is said to be complete. we define a unique f f P P risk-neutral probability measure I . .1) For 5. The market price of risk is an adapted process (t) = ( 1 (t) satisfying the system of equations (MPR) above. Case II: (No solution. Martingale Representation Theorem Discounted stock prices 201 d X d Si((tt)) = ( i(t){z r(t)) Si((tt)) dt + Si((tt)) . (MPR) has a unique solution (t). The market admits no arbitrage. (Harrison and Pliska. A stochastic calculus model of continuous trading: complete markets.): If a market has a risk-neutral probability measure. the discounted wealth process corresponding to any portfolio process is a I P martingale. the Martingale Representation Theorem can be used to show that every contingent claim can be hedged.1 to be satisfied. then there are multiple risk-neutral probability measures. (Harrison and Pliska. Case III: (Multiple solutions). and this implies that the market admits no arbitrage. but there are contingent claims which cannot be hedged. and Applications 15 (1983).) If (MPR) has no solution. Stochastic Proc.60 (Fundamental Theorem of Asset Pricing) Part I. There are three cases to consider: d (t)) Case I: (Unique Solution). For Lebesgue-almost every t and IP -almost every ! .

202 .

1 1 2 2 2. 2) 2 dB2 dB2 2 2 2 S 2 dt = 2 2 dS1 dS2 = S1 1S2 2 dB1 dB1 = 1 2S1S2 dt: In other words. To F (t) 0 dS1 = S1 dS2 = S2 We assume 1 1 dt + 1 dB1] q 2 dt + 2 dB1 + 1 . t T F P). all processes can depend on t and ! .Chapter 19 A two-dimensional market model Let B (t) In what follows. 1 S1 dS2 has instantaneous variance 2 . we omit the arguments whenever there is no ambiguity. dS1 has instantaneous variance 2 . 1: Note that 2 2 dB2 : >0 2>0 .r 203 . (t) = exp Zt 0 r du : = 2= 1. but are adapted to simplify notation. 2 S2 dS1 and dS2 have instantaneous covariance S1 S2 Accumulation factor: 1 2.1 2 2 dS1 dS2 = S1 2 dB1 dB1 = 2S1 dt 1 1 2 2 dS2 dS2 = S2 2 2 dB1 dB1 + S2 (1 . Stocks: F (t) 0 t T = (B1 (t) B2(t)) 0 be a two-dimensional Brownian motion on ( be the filtration generated by B . Let t T.r (MPR) The market price of risk equations are 2 1+ q 1.

1 ( 1 + 2 ) du 2 0 0 0 8A 2 F : f IP is the unique risk-neutral measure. r) . d X = 1 (dX . rS2 dt) e 1 2S2 1S1 1 dB1 + f V Y (t) = IE (T ) F (t) q e1 + 1 .1 < < 1. 19. 2 Z f(A) = Z (T ) dIP IP A Z (t) = exp 1 2).1 Hedging when .204 The solution to these equations is 1 2 provided . Suppose . 1 1 2 1. r) p 2( 1 . 2S2 ) dt dX = 1 dS1 + 2 dS2 + r(X . 1 dB1 . Define the I -martingale . Define e B1 (t) = e B2(t) = Then Zt Z0 0 1 du + B1 (t) 2 du + B2 (t): t dS1 = S1 r dt + dS2 = S2 r dt + h e 1 dB1 i q 2 2 dB2 e e 2 dB1 + 1 .r 1 ( 2 . rX dt) =1 =1 1(dS1 . we define Zt Zt Zt 2 2 . : We have changed the mean rates of return of the stock prices. rS1 dt) + 1 2 (dS2 . but not the variances and covariances. 2 dB2 . 2 dB 0 t T: 2 2 e dB2 : f P Let V be F (T )-measurable.1 < = = < 1.1 < <1 1S1 . Then (MPR) has a unique solution ( 1.

X (T ) = V: Every F (T )-measurable random variable can be hedged. 19.2 Hedging when The case =1 = . the market is complete.1 is analogous. There are two cases: (MPR) . The market price of risk equations are 1 dt + 1 dB1 ] 2 dt + 2 dB1 ] 1 1= 1. A two-dimensional market model The Martingale Representation Corollary implies 205 Y (t) = Y (0) + We have Zt 0 e 1 dB1 + Zt 0 e 2 dB2: e dB1 d X = 1 1 S1 1 + 1 2 S2 dY = We solve the equations e e 1 dB1 + 2 dB2 : 1 S1 1 + +1 2S2 q 2 1. With this choice of ( 1 f V X (0) = Y (0) = IE (T ) we have X (t) = Y (t) 0 t T and in particular.CHAPTER 19.r 2 1= 2.r The process 2 is free. Then dS1 = S1 dS2 = S2 The stocks are perfectly correlated. q 2 S2 2= 1 2 2= 2 2) and setting 2). Assume that = 1. 2 2 dB2 e 1 1 1 for the hedging portfolio ( 1 2S2 1 .

1 | 1 . 2 2 Then d X =1 =1 1. 2 r : Set 1= 1 S1 2=. rS1 dt) + 1 2(dS2 . r) dt + 1 2 S2 2 1 dt + dB1 ] 2 dB1] 1S1 1 1 dt + dB1 ] + 1 2S2 2 e dB1 : e Notice that B2 does not appear. if and 1 or 2 depend on B2 . Let I be one of them. This market admits arbitrage.r 6= 2 . For example.r : There is no solution to (MPR). 2 r : The market price of risk equations 1 1 2 1 = = 1. r) dt + 1 1 dB1] + 1 2S2 ( 2 . If V depends on B2 . then there is trouble. then it can probably not be hedged. there are infinitely many risk-neutral measures. r dt 1 {z 2 } Positive Case II: r = 2. r) dt + 1 : S 2 . 1 1 1 1 . 2 . Hedging: 1.r 1 = 2. Indeed d X = 1 1(dS1 . r) dt + =1 = 1 1 S1 1 + 1 dB1] + 1 2S2 ( 2 . r dt + dB . r .206 Case I: 1 . 1 1S1 ( 1 . Let V be an F (T )-measurable random variable. r dt + dB 1 2 2 dB1] r > 2.r 2. V = h(S1 (T ) S2(T )) . rS2 dt) =1 Suppose 1. there is no risk-neutral 1 2 measure. and consequently.r have the solution 1= f P 2 is free.r 2 d X = 1 1S1 ( 1 .

A two-dimensional market model 207 f P More precisely. we must have 2 = 0: .CHAPTER 19. we define the I -martingale f V Y (t) = IE (T ) F (t) We can write 0 t T: Y (t) = Y (0) + so Zt 0 e 1 dB1 + Zt 0 e 2 dB2 dY = e e 1 dB1 + 2 dB2 : To get d X to match dY .

208 .

b) exp . ( ) ! With drift. b)2 dm db m > 0 b < m: m = 2T T 2 T 1 .Chapter 20 Pricing Exotic Options 20.1 Reflection principle for Brownian motion Without drift. @m 2T 2 T ( ) 2(2p . (2m . @m @b p 1 exp . b)2 dm db = . bg ( 2) 1 Z x =p exp . m>0 b<m So the joint density is Z 2 @2 IP fM (T ) 2 dm B(T ) 2 dbg = . (2m . x dx dm db 2T 2 T 2m b ( )! @ p 1 exp . 2T dx 2 T 2m b 1 . Define M (T ) = 0maxT B (t): t Then we have: IP fM (T ) > m B(T ) < bg = IP fB (T ) > 2m . Let e B(t) = t + B(t) 209 .

2: Possible values of B (T ) M (T ). M(T)) lies in here b Figure 20.210 2m-b shadow path m b Brownian motion Figure 20. .1: Reflection Principle for Brownian motion without drift m m=b (B(T).

we conclude that (MPR) f IP fM (T ) 2 dm B (T ) 2 d~g ~ e b 1 2 T g I fM (T ) 2 dm B (T ) 2 d~g f f e = expf ~ . b) exp . ~) exp . 1 2 T g 2 = m= ~Z 1 h i ~=m bZ ~ . 2 2 T g IP fM (T ) 2 dm B (T ) 2 d~g: ~ b b 1 ~ e b But also.1 m=0 ~= ~ b f f h(m ~) expf ~ . B 2 Z f(A) = Z (T ) dIP 8A 2 F : IP A f SetM (T ) = max0 t T f Under I P e B (T ): ) ( m b ~ ~ b ffM (T ) 2 dm B (T ) 2 d~g = 2(2p . 2 b b )~ (P m ~ ~ ~ ~2 b 2 = 2(2p . 1 2 T gdm d~ m > 0 ~ < m: ~ b ~ b ~ 2T T 2 T . ~)2 dm d~ m > 0 ~ < m: f e ~ b ~ b ~ IP ~ b 2T T 2 T ~) be a function of two variables. so Let h(m ~ T f e f IEh(M (T ) B (T )) = IE h(M (Z ()TB (T )) ) f e f f e e = IE h(M (T ) B (T )) expf B (T ) . Define Z (T ) = expf. b) : expf ~ .1 m=0 ~= ~ b f h(m ~) IP fM (T ) 2 dm B (T ) 2 d~g: ~ b ~ e b Since h is arbitrary. 1 2 T g 2 = expf.CHAPTER 20. (2m . Then b e B is a Brownian motion (without drift). (2m . f e IEh(M (T ) B (T )) = m= ~Z 1 ~Z m b= ~ . B (T ) . (B (T ) + T ) + 1 2 T g 2 e (t) + 1 2T g = expf. is a Brownian motion (without drift) on ( F P). Pricing Exotic Options where B (t) 211 0 t T .

K 1 . e = e rT IE S (0) expf B (T )g .212 20. K )+ 1 S (T )<L f g To simplify notation.2 e B (t) = t + B(t): Consequently. 6 4 | {z 2 39 > 7> = t7> 7 } 5> where = r. rT IE h i + f g . K )+ 1 S (T )<L e = e rT IE S (0) expf B (T )g . dS (t) = rS (t) dt + S (t) dB (t) S (t) = S0 expf B (t) + (r . K )+ 1 S (T )<L : Because IP is the risk-neutral measure. v (0 S (0)) = e (S (T ) . f g h i 8 > > < = S0 exp > > : e = S0 expf B (t)g 2 6 6B(t) + r . f S (t) = S0 expf M (t)g where. " 1 S(0)exp M (T ) e f f g <L g e B (T )> 1 log K M (T )< 1 log L e S (0)} | {z | {zS (0)} ~ b # m ~ . The payoff at time T is (S (T ) . K . Let 0 < K where < L be given.2 Up and out European call. assume that IP is already the risk-neutral measure. so the value at time zero of the option is S (T ) = 0maxT S (t): t v (0 S (0)) = e rT IE (S (T ) . 1 2 )tg 2 . f e M (t) = 0maxt B (u): u We compute.

(2y . 1 2 T dx . 1 2 T y=m dx rT = .3: Possible values of B (T ) We consider only the case f M (T ). Z mZ m ~ ~ ( ) The standard method for all these integrals is to complete the square in the exponent and then recognize a cumulative normal distribution. 1 2 T dx ~ .CHAPTER 20. 1 2 T dx ~ . Pricing Exotic Options 213 ~ M(T) y ~ m (B(T). x2 + x . 1 2 T dx: ~ +p 2 ~ 2T 2 T b rT . . K ) p 1 exp . 2 2 T dy dx ~ x b T 2 T( ) ~ Zm ~ 1 exp . x) exp . 1 2 T 2 2T b 2 T )# ( (2m . exp . 2T 2 ( ) ~ 1 e rT S (0) Z m exp x . M(T)) lies in here x ~ B(T) x=y ~ b e Figure 20. x)2 + x .e (S (0) expf xg . x + x . (2m . . 2 1 p v (0 S (0)) = e (S (0) expf xg . (2m . 1 2 T dx =p 2 ~ 2T 2 T b ) ( ~ 1 e rT K Z m exp . b ~ R m R m : : :dy dx: ~ ~ We compute ~ x b . . x)2 + x . K < S (0) L leads to ~ < 0 m and the analysis is similar. . . p2 T 2 ~ 2T b ( ) ~ 1 e rT S (0) Z m exp x . x)2 + x . x)2 + x . S (0) K < L so 0 ~ < m: b ~ The other case. (2y2T x) + x . . K ) p 2 ~ 2T 2 T b y=x " ( 2 ) Zm ~ = e rT ~ (S (0) expf xg . K ) 2(2y . x2 + x .p 2 ~ 2T 2 T b ) ( ~ 1 e rT K Z m exp . We carry out the details for the first integral and just .

r T + T m ~ rT K N p . The exponent in the first integrand is = . T 2 dx =p ~ 2T 2 2 T b p m p p~T .214 give the result for the other three. 2 T T . ~) + r T + T ~ p . r T . r T . 21 x . 2 T T " T p p ! p p !# ~ T . 2T + x . r T . 2 T T T Putting all four integrals together. p p ! T ~ ~ p N (2m . 2 + rT: T In the first integral we make the change of variable x2 x . we have " p p pT . r T . rT= . N pb . 2T + x . 2 T . N (2m . rT . 2 T expf. N pb .rT + 2m r . r T + . 1 2T dx 2 ( ) ~ 1 S (0) Z m exp . y2 g dy 2 p p ! p p !# : ~ m ~ v (0 S (0)) = S (0) N p . 2 T . T . S (0) N pT 2 2 pT p ! m ~ + exp . 1 x . T )2 + 1 2T + T 2 T 2 T = . N pb . 21 (x . 2 T ( ) = p 1 S (0): 2 T ~ b Z ~ m ~ = S (0) N p . T=2)= T dy = dx= T to obtain p p rT ~ e p S (0) Z m exp b 2 T ~ .e 2 2 T T " p p ! p p !# ~ b m + r T + T . x2 x . rT . " p p ! p p !# where ~ = 1 log K b S (0) m = 1 log SL : ~ (0) . 1 2 T 2 y = (x . r T . 2 ~ N p + r T . r T . 2T . b) + r T .

Pricing Exotic Options 215 L v(t. . " ~ b p p !# If we replace T by T . r T + . N p . f S (T )<L i g 0 t T 0 x L: This function satisfies the terminal condition v(T x) = (x . . We consider the function v(t x) = IE t x e h . but a similar albeit longer formula can also be derived for K < x L. K )+ 1 . 2 T T " p p !# ~ b T rT K 1 .K)+ v(t.e K 2 T . t and replace S (0) by x in the formula for v (0 S (0)).e 2 T p p ! 1 (0) = S (0)N p log SK + r T + 2 T T p p ! T : 1 rT KN p log S (0) + r T . We have actually derived the formula under the assumption x K L. the value of the option at the time t if S (t) = x.L) = 0 v(T.4: Initial and boundary conditions. N p .x) = (x .0) = 0 T Figure 20.rv + vt + rxvx + 1 2x2vxx 0 t < T 0 x L: 2 . If we let L!1 we obtain the classical Black-Scholes formula v (0 S (0)) = S (0) 1 .CHAPTER 20. we obtain a formula for v (t x). K )+ and the boundary conditions 0 x<L v(t 0) = 0 0 t T v(t L) = 0 0 t T: We show that v satisfies the Black-Scholes equation . r T . r(T t) (S (T ) .

then IE e But when . K )+ 1 S (T )<L F (t) F (u) = IE e rT (S (T ) . . Proof: First note that . K )+ 1 S (T )<L F (t) (! ) i rT (S (T ) . K )+ 1 S (T )<L F (u) = e r(u )v (u ^ S (u ^ )) : . K )+ 1 f S (T )<L g F (t) (!) = 0: (!) t. rT (S (T ) . f g . = IE e = e rt v (t S (t !)) = e r(t (!))v (t ^ . and choose t 2 0 T ]. if (!) > t. we have v (t ^ (!) S (t ^ (! ) !)) = v(t ^ (!) L) = 0 so we may write IE e . ^ h t S (t !) rT (S (T ) . rT (S (T ) . then the Markov property implies IE e . we have e Suppose 0 .216 Let S (0) > 0 be given and define the stopping time = minft 0 S (t) = Lg: Theorem 2. In both cases. S (t ^ )) F (u) f g = IE IE e rT (S (T ) . K )+ 1 S (T )<L f g f g S (t ^ (!) !)) : . Then IE e . r(t (!)) v (t ^ ^ (!) S (t ^ (! ) !)) : On the other hand. If (!) t. .61 The process e is a martingale. r(t )v (t ^ ^ S (t ^ )) 0 t T S (T ) < L () > T: Let ! 2 be given. K )+ 1 f S (T )<L g F (t) (!) = e . ^ . r(t )v (t ^ ^ S (t ^ )) = IE e r(t )v (t ^ ^ . rT (S (T ) . K )+ 1 f S (T )<L g F (t) : u t T .

ru (. S (t)vx(t S (t)) dB (t) 0 t Let X (t) be the wealth process corresponding to some portfolio (t). rt : rt (t) S (t) dB (t): X (0) = v (0 S (0)) (t) = vx (t S (t)) 0 t T^ : X (T ^ ) = v(T ^ S (T ^ )) ( v (T S (T )) = (S (T ) . . Therefore. . K )+ = v ( L) = 0 if if >T T. 0 + Because e. .rv + vt + rxvx + 1 2x2vxx = 0 0 t T 0 x L 2 d e rtv (t S (t)) = e .rv + vt + rSvx + 1 2 S 2vxx) du 2 ^ .rv + v 2 2 t + rSvx + 1 S vxx ) du 2 is a martingale.r(t^ )v (t ^ Zt |0 ^ e . rt Svx dB: Integrate from 0 to t ^ : e . The Hedge . r(t )v (t ^ ^ S (t ^ )) = v(0 S (0)) Zt + e ru (. Pricing Exotic Options 217 For 0 t T .CHAPTER 20. the Riemann integral ^ e . Then d(e rtX (t)) = e We should take and Then . we compute the differential d e rt v(t S (t)) = e rt(.rv(u S (u))+ vt(u S (u)) + rS (u)vx(u S (u))+ 2 2S 2(u)vxx(u S (u)) = 0 0 u t ^ : The PDE then follows. . 1 .rv + vt + rSvx + 1 2S 2vxx ) dt + e 2 . ru A stopped martingale is still a martingale {z Svx dB: } Zt 0 S (t ^ )) is also a martingale. .

Rather than using the boundary condition v(t L) = 0 0 t T solve the PDE with the boundary condition where short position. and consequently he is left with no money. pays off the option. and is left with zero. a small error in hedging can create a significant effect. x) 0 K L x Figure 20.5. the option has “knocked out”. x) 0 K L x v(t. 20. At the boundary.218 v(T. times the dollar 2. he is now in a region of (t) = v x(t S (t)) near zero. v (t x) has the form shown in the bottom part of Fig. Because a large short position is being taken. (t) = vx (t S (t)) In particular.3 A practical issue For t < T but t near T . . he covers this short position with funds from the money market.5: Practial issue. The value of the portfolio is always sufficient to cover a hedging error of size of the short position. 20. This is more expensive than before. However. The new boundary condition guarantees: 1. the hedging portfolio can become very negative near the knockout boundary. because the stock price has risen. The hedger is in an unstable situation. v(t L) + Lvx (t L) = 0 0 t T is a “tolerance parameter”. say 1%. He should take a large short position in the stock. Here is a possible resolution. He should cover his short position with the money market. Lvx (t L) is the dollar size of the Lvx (t L) remains bounded. If the stock does not cross the barrier L. so no money is needed to pay it off. If the stock moves across the barrier.

B(t)) + (r . rT h ZT 0 S (t) dt : !# Introduce an auxiliary process Y (t) by specifying dY (t) = S (t) dt: With the initial conditions S (t) = x Y (t) = y we have the solutions S (T ) = x exp (B(T ) . t) 2 Y (T ) = y + n o ZT t S (u) du: Define the undiscounted expected payoff u(t x y ) = IE t x y h(Y (T )) 0 t T x 0 y 2 IR: 219 .Chapter 21 Asian Options Stock: dS (t) = rS (t) dt + S (t) dB (t): V =h Payoff: ZT 0 S (t) dt ! Value of the payoff at time zero: X (0) = IE e " . 1 2 )(T .

220 21. . . where The PDE for v is Zt 0 . S (u) du .rv + vt + rxvx + 1 2x2vxx + xvy = 0 2 21. The differential of the value X (t) of a portfolio dX = dS + r(X . v (t x y ) = e r(T t) u(t x y ): (1.1 Feynman-Kac Theorem The function u satisfies the PDE 1 ut + rxux + 2 2x2 uxx + xuy = 0 0 t T x 0 y 2 IR the terminal condition and the boundary condition u(T x y ) = h(y ) x 0 y 2 IR u(t 0 y) = h(y ) 0 t T y 2 IR: One can solve this equation.2 Constructing the hedge Start with the stock price S (0).1) v (T x y ) = h(y ) v(t 0 y) = e r(T t)h(y ): One can solve this equation rather than the equation for u. S ) dt = S (r dt + dB ) + rX dt . . Then v t S (t) is the option value at time t. r S dt = S dB + rX dt: (t) is We want to have X (t) = v t S (t) so that Zt 0 S (u) du S (u) du X (T ) = v T S (0) =h ZT 0 ! ZT 0 S (u) du : ! .

S ( ) 0): x 0 . This value is The function w satisfies the Black-Scholes PDE w(t x) = IE t xe . starting from the same initial condition. 0 t T: 0 t 8 <w (t S (t)) R (t) = : x vx t S (t) t S (u) du < t T: . ! < T .CHAPTER 21.rw + wt + rxwx + 1 2x2wxx = 0 0 t 2 w( x) = v ( x 0) x 0 w(t 0) = e .1) Compare this with Take dX (t) = rX (t) dt + (t) S (t) dB(t): (t) = vx (t S (t)): If X (0) = v (0 S (0) 0). Asian Options The differential of the value of the option is 221 dv t S (t) Zt 0 S (u) du = vtdt + vxdS + vy S dt + 1 vxx dS dS 2 1 = (vt + rSvx + Svy + 2 2 S 2vxx ) dt + Svx dB = rv (t S (t)) dt + vx (t S (t)) S (t) dB (t): (From Eq. r( t) v ( . 21. then X (t) = v t S (t) Zt 0 S (u) du 0 t T because both these processes satisfy the same stochastic differential equation. we compute next the value of a derivative security which pays off v ( S ( ) 0) at time . . with terminal condition and boundary condition The hedge is given by r(T t)h(0) . We compute v( x y ) = IE x y e r(T )h(Y (T )) just as before. For 0 t .3 Partial average payoff Asian option Now suppose the payoff is V =h where 0 < ZT S (t) dt . 1.

Asian options.222 Remark 21. Geman and M. Bessel processes. t)) has been computed.1 While no closed-form for the Asian option price is known. . See H. Math. Finance 3 (1993). the Laplace transform (in 2 the variable 4 (T . Yor. 349–375. and perpetuities.

Chapter 22 Summary of Arbitrage Pricing Theory A simple European derivative security makes a random payment at a time fixed in advance. We have no probabilities at this point. 22. The value at time t of such a security is the amount of wealth needed at time t in order to replicate the security by trading in the market. k Sk ): Given a simple European derivative security V (! 1 use a portfolio processes !2). 2. Solving the equations. we obtain: 223 . Hedging Portfolio Let be the set of all possible sequences of n coin-tosses. The hedging portfolio is a specification of how to do this trading.1 Binomial model.1) Evolution of the value of a portfolio: Xk+1 = k Sk+1 + (1 + r)(Xk . (See Fig. Let r 0 u > r + 1 d = 1=u be given. we want to start with a nonrandom X0 and 1 (T ) 0 so that 1 (H ) X2(!1 !2 ) = V (!1 !2 ) 8! 1 !2 : There are four unknowns: X0 (four equations) 0 1(H ) 1(T ).

k=0 (log Sk+1 . From a practical point of view.e. (1 + r) X1(T ) u d u. Sk : k In continuous time. X2(!1 T ) 1 (!1 ) = S (! H ) . then we change . We reiterate that the probabilities are only introduced as an aid to understanding and computation.. . Xk+1 . log Sk )2 = log SS+1 k = ( log u)2 = (log u)2 : Let 2 = log u > 0. We want to find a description of the paths in the model. because we have a hedge which works on every path. d | {z }5 r 1 X0 = 1 + r 1 + . . (1 + r) X2(! 1 T )7 4 u d | {z ) u . we will have the analogous equation (t) d X((tt)) = (t) d S(t) : . what matters is that the paths in the model include all the possibilities.224 2 3 7 1 6 r X1(!1) = 1 + r 6 1 + . They all have the property k (log Sk+1 . S (! T ) 2 1 2 1 X1(H ) . k Sk ): k Define Then = (1 + r)k : Xk+1 = k+1 k+1 k Sk+1 + Xk . d X2(!1 H } + u . Recall: Xk+1 = k Sk+1 + (1 + r)(Xk . d X1(H ) + u . Then n 1 X .d X2(!1 H ) . k+1 k k k+1 k Sk k i. The paths of log Sk accumulate quadratic variation at rate 2 per unit time. log Sk )2 = 2 n: If we change u. X1(T ) : 0 = S (H ) . Xk = k Sk+1 . and the pricing and hedging formulas on the previous page will give different results. S (T ) 1 1 V (!1 H ) V (! 1 T ) The probabilities of the stock price paths are irrelevant.

Using the binomial model as a guide. Summary of Arbitrage Pricing Theory 225 k f If we introduce a probability measure I under which Sk is a martingale. However. q = IP f!k = T g. we use Brownian motion. we need a probability measure.. we denote p= ~ r p = 1 + . the only thing about this probability measure which ultimately matters is the set of paths to which it assigns probability zero. Sk : Xk + f E = k I k k | k+1 =0 {z k } Then we Suppose we want to have must have X2 = V . then P martingale. is a continuous function of t. To generate these paths.CHAPTER 22.2 Setting up the continuous model Now the stock price S (t) 0 t T . Sk F k k+1 k k+1 k Sk+1 F . (1 + r) : ~ u. Indeed.d 22. These are the paths with volatility 2. we choose > 0 and try to hedge along every path S (t) for which the quadratic variation of log S (t) accumulates at rate 2 per unit time. d ~ u d q = u . Xk k will also be a f f IE Xk+1 F k = IE Xk + k Sk+1 . but that is impossible. To introduce Brownian motion. regardless of the portfolio used. and compute ~ f f ~ ~ IE Sk+1 F k = pu Sk + qd Sk k+1 k+1 k+1 1 pu + q d] Sk : = 1+r ~ ~ k We need to choose p and q so that ~ ~ pu + qd = 1 + r ~ ~ p + q = 1: ~ ~ The solution of these equations is f IP under which Skk is a martingale. rather than coin-tossing. where V is some F 2-measurable random variable. We would like to hedge along every possible path of S (t). . X1 f X2 1 f V 1 + r X1 = 1 = IE 2 F 1 = IE 2 F 1 f f X0 = X0 = IE X1 = IE V : 0 1 2 To find the risk-neutral probability measure f IP f!k = H g.

the choice of in this definition is irrelevant. 1t + B (t !1) 0 t T B(t !1 ) is a different function. the paths of t + B(t) S (t) = S (0) expf t + B(t)g For any accumulate quadratic variation at rate 2 per unit time. Then. 1 2tg 2 e rt S (t) = S (0) expf B(t) . We want to define so that the paths of log S (t) = log S (0) + t + B (t) accumulate quadratic variation at rate 2 per unit time. there is an ! 2 2 such that 1t + B (t !1 ) = 2t + B (t !2) 0 t T: In other words. Surprisingly. is a continuous function of t. the reason for this is the following: Choose ! 1 2 . be a Brownian motion defined on a probability space ( F P). The mathematically precise statement is the following: If a set of stock price paths has a positive probability when S (t) is defined by S (t) = S (0) expf 1t + B(t)g then this set of paths has positive probability when S (t) is defined by S (t) = S (0) expf 2t + B(t)g: Since we are interested in hedging along every path. the choice of is irrelevant. is a martingale under IP .226 Let B (t) 0 t T . 2 2 S (t) = S (0) expfrt + B(t) . Roughly. for 1 2 IR. except possibly for a set of paths which has probability zero. With this choice of . regardless of whether we use 1 or 2 in the definition of S (t). 1 2tg 2 . dS (t) = rS (t) dt + S (t) dB (t) . then 2 t + However. The most convenient choice of so and is 1 = r. 2 IR. we will see the same paths. If we replace 1 by 2.

we must have value X (0) so that X (T ) = V . and initial portfolio X (t) must be a martingale. 2 e B has the same paths as B . the payoff of a simple European derivative security.1) (3. Summary of Arbitrage Pricing Theory and IP is the risk-neutral measure. r dt + dB (t) f We can change to the risk-neutral measure I . 22. Then e dS (t) = rS (t) dt + S (t) dB (t) e f P where B is a Brownian motion under I . X((tt)) is a martingale under I . Because (t) X (t) = IE V F (t) f (t) (T ) 0 t T: (3. Set (t) = ert : Then f P so S (t) is a martingale under I .3) This is the risk-neutral pricing formula.3 Risk-neutral pricing and hedging {z e dB (t) i } e B is a : f P Let I denote the risk-neutral measure. we have 227 S (t) = S (0) expf t + B(t)g dS (t) = ( +{z1 2 ) S (t) dt + S (t) dB(t): | 2 } = rS (t) dt + h | . If a different choice of is made. under which P Brownian motion. (t) Evolution of the value of a portfolio: which is equivalent to (t) (t) e d S(t) = S(t) dB (t) dX (t) = (t)dS (t) + r(X (t) . P Now suppose V is a given F (T )-measurable random variable. 1 2 . (t)S (t)) dt (t) d X((tt)) = (t)d S(t) (t) e = (t) S(t) dB (t): (3.CHAPTER 22. We have the following sequence: .2) f Regardless of the portfolio used. We want to find the portfolio process (T ) 0 t T . and then proceed as if had been chosen to be equal to r .

the risk-neutral pricing formula is valid provided there is a unique risk-neutral measure.3). 3. and (3. is the value of f V X (0) = IE (T ) and it will end with value V f V X (T ) = (T )IE (T ) F (T ) = (T ) (T ) = V: Remark 22. the risk-neutral pricing formula is still “valid” when r and are processes adapted to the filtration generated by B .1) or (3. If they depend on e either B or on S . because we do not yet have (t)). If you start with then there is a hedging portfolio 2. (t) so that (3.2). X (t) = IE V F (t) : f (t) (T ) Remark 22. Comparing (3.2). To carry out step 3.4) implies (3. which implies (3. they are adapted to the filtration generated by B . we first use the tower property to show that X((tt)) defined by (3. when there are multiple assets and/or multiple Brownian motions. 0 t T . The “validity” of the risk-neutral pricing formula means: 1. such that X (T ) = V . 2.2).4). We next use the corollary to the Martingale Representation Theorem (Homework Problem 4. the value X (t) of the hedging portfolio in 1 satisfies f V X (0) = IE (T ) (t) 0 t T . A probability measure is said to be risk-neutral provided .228 1. which we know.3) is a martingale f P under I . (3. Construct defined in step 2.5) 0 t T .4) for some proecss . which we want. At each time t. we see that the hedging portfolio must begin with value Then (3. by (3.1 Although we have taken r and to be constant.2) (or equivalently. (3.1).2 In general.5) to show that e d X((tt)) = (t) dB (t) (t) = (t) (t) : S (t) From (3. Define X (t) V is given.1)) is satisfied by the X (t) 0 t T . we decide to define (3. the definition of X .3) (not by (3. which implies that X (t) the portfolio process (t) 0 t T .

1 Assume r and variable.t) v(t x y) = IE e Y (T) = y + h(Y (T)) i where ZT t S(u) du: . 229 To see if the risk-neutral measure is unique. We can take the stock price to be the state i e h v(t x) = IE t x e. Summary of Arbitrage Pricing Theory it has the same probability-zero sets as the original measure. it makes all the discounted asset prices be martingales. so that f V X (t) = (t)IE (T ) F (t) is a function of these variables. Define are constant. are constant. Example 22. and V = h(S(T)).rT h(S(T)) F (t) = v(t S(t)) (t) and X(t) e = e.r(T .r(T .4 Implementation of risk-neutral pricing and hedging To get a computable result from the general risk-neutral pricing formula X (t) = IE V F (t) f (t) (T ) one uses the Markov property. compute the differential of all discounted asset prices e e and check if there is more than one way to define B so that all these differentials have only d B terms. We need to identify some state variables. 22.CHAPTER 22.rt v(t S(t)) is a martingale under IP . Define e t x y h . Example 22. the stock price and possibly other variables.2 Assume r and V =h Take S(t) and Y (t) = 0 ZT 0 S(u) du : ! R t S(u) du to be the state variables.t) h(S(T)) : Then e X(t) = ert IE e.

We take the differential and set the dt term to zero. X (t) = X (0) + Z t (u) S (u) dB (u) e (t) (u) 0 . and this equation must e hold wherever the state processes can be. we get an expression involving v to be a martingale. This gives us a partial differential equation for v .4) to obtain (3. e X(t) = e.rT h(S(T)) F (t) = v(t S(t) Y (t)) and is a martingale under IP . Z t r(u Y (u)) du v(t S(t) Y (t)) (t) 0 In every case. Define 2 3 6 ( Z 7 ) 6 7 T txy6 e 6exp . The dB term in the differential of the equation is the differential of a martingale. r(u Y (u)) du h(S(T ))7 : 7 v(t x y) = IE 6 7 t 6| 7 4 5 {z } (t) (T ) Then e X(t) = (t)IE h(S(T)) F (t) " ( (T) e = IE exp . e X(t) = exp .5).3 (Homework problem 4. ZT t r(u Y (u)) du h(S(T )) F (t) ) # = v(t S(t) Y (t)) and is a martingale under IP . and since the martingale is we can solve for (t).rt v(t S(t) Y (t)) (t) Example 22.230 Then e X(t) = ert IE e.2) e dS(t) = r(t Y (t)) S(t)dt + (t Y (t))S(t) dB(t) e dY (t) = (t Y (t)) dt + (t Y (t)) dB(t) V = h(S(T)): Take S(t) and Y (t) to be the state variables. This is the argument which uses (3.

rv + vt + rSvx + vy + 1 2 S 2 vxx + Svxy + 2 2 vyy ) dt 2 e + ( Svx + vy ) dB The partial differential equation satisfied by v is . and all other variables are functions of (t y).rv + vt + rxvx + vy + 1 2x2 vxx + 2 where it should be noted that v xvxy + 1 2 vyy = 0 2 where (see (3.2)) = v(t x y). = (t Y (t)). We have e 1 d X(t) = (t) . Summary of Arbitrage Pricing Theory Example 22. Z t r(u Y (u)) du v(t S(t) Y (t)) (t) | 0 {z } 1= (t) is a martingale under IP . We have 1 e d X(t) = (t) Svx + vy ] dB(t) (t) = (t Y (t)).CHAPTER 22. we should take Y (t)) (t) = vx (t S(t) Y (t)) + (t (t (t)) S(t) vy (t S(t) Y (t)): Y . and S = S(t). v = v(t S(t) Y (t)).r(t Y (t))v(t S(t) Y (t)) dt (t) + vt dt + vx dS + vy dY + 1 vxx dS dS + vxy dS dY + 1 vyy dY dY 2 2 1 1 = (t) (.4 (Continuation of Example 22.3) 231 X(t) = exp . We want to choose (t) so that (t) to be e d X(t) = (t) (t Y (t)) S(t) dB(t): (t) (t) Therefore.

232 .

.62 (Levy) Let B (t) F (t) 0 t T . We compute (this uses the continuity condition (1) of the theorem) deuB(t) = ueuB(t)dB(t) + 1 u2 euB(t)dB(t) dB(t) 2 so euB (t) = euB (s) + Zt s ueuB (v) dB(v ) + 1 u2 2 233 Zt s euB(v) uses cond. s.. 3 dv: |{z} . B (s)) F (s) 1 2 = e 2 u (t s) : . B(s) is independent of F (s). i. hBi(t) = t 0 t T . We need to show that B (t) . Since the moment generating function characterizes the distribution. (i. such that: 0 t T be a process on ( F P).Chapter 23 Recognizing a Brownian Motion Theorem 0. B is a martingale. adapted to a filtration 1.e. Then B is a Brownian motion. B (s) is independent of F (s). B (s) is normal. 2. s. with mean zero and variance t . 3. Proof: (Idea) Let 0 s < t T be given. B (s) is IE eu(B (t) . B (s) is normal with mean 0 and variance t . informally dB(t) dB(t) = dt). B (t) . and B (t) . and conditioning on F (s) does not affect this. We shall show that the conditional moment generating function of B (t) . this shows that B (t) .e. the paths of B (t) are continuous.

1 u2 s 2 : IE euB(t) F (s) = '(t) = euB(s)+ 2 u2 (t IE eu(B(t) . s) B (s)) F (s) 1 2 = e 2 u ( t s) : . and so IE Zt s =.234 Rt Now 0 ueuB (v)dB (v ) is a martingale (by condition 2). 1 . we get euB (s) = ke 2 u2 s =)k = euB (s) Therefore. 1 . = 0: It follows that Zs 0 ueuB(v) dB (v ) F (s) ueuB (v) dB(v ) + IE Zt 0 ueuB(v)dB (v ) F (s) IE euB(t) F (s) = euB(s) + 1 u2 2 We define Zt s IE euB(v) F (s) dv: '(v) = IE euB(v) F (s) so that '(s) = euB(s) and 1 '(t) = euB(s) + 2 u2 1 ' (t) = 2 u2'(t) 1 '(t) = ke 2 u2 t : 0 Zt s '(v) dv Plugging in s. .

Indeed.1 Identifying volatility and correlation Let B1 and B2 be independent Brownian motions and dS1 = r dt + S1 dS2 = r dt + S 2 11 dB1 + 12 dB2 21 dB1 + 22 dB2 Define 2 + 2 1= q 11 12 2 + 2 2= 21 22 11 21 + 12 = Define processes W1 and W2 by q 1 2 22 : dW1 = dW2 = Then W1 and W2 have continuous paths. dS1 = r dt + S1 dS2 = r dt + S 2 1 dW1 2 dW2: 12 dB2 )( 21dB1 + 22dB2 ) dW1 dW2 = 1 ( 11dB1 + = 1 ( 1 2 = dt: 1 2 11 21 + 12 22 ) dt . and 11 dB1 + 12 dB2 1 dB1 + 22 dB2 : 21 2 dW1 dW1 = 12 ( 11dB1 + = 12 ( = dt 1 1 2 dB 11 1 12 dB2 )2 2 dB dB ) 12 2 2 dB1 + and similarly Therefore. are martingales. The stock prices have the representation dW2 dW2 = dt: The Brownian motions W1 and W2 are correlated.CHAPTER 23. W1 and W2 are Brownian motions. Recognizing a Brownian Motion 235 23.

2 2 dB2 ( 6= 1) dB1 dB1 = dW1 dW1 = dt 1 dB2 dB2 = 1 . We want to find dS1 = r dt + dW 1 1 S1 dS2 = r dt + dW 2 2 S 2 = so that 0 " 11 21 11 12 12 22 21 22 # = = = " " " 11 21 2 + 2 11 12 + 12 22 11 21 # 2 1 2 1 2 1 2 2 11 = 1 12 22 #" # # 11 21 + 12 22 2 + 2 21 22 A simple (but not unique) solution is (see Chapter 19) 12 = 0 22 = 21 = This corresponds to 2 q 1. dW1 1. we have 1 dW1 = 1 dB1 =)dB1 = dW1 q 2 dW2 = 2 dB1 + 1 . = dt 2 dt . 2 2 dt + 2 dt . then there is no B2 and dW2 = dB1 = dW1: Continuing in the case 6= 1. If = 1. 2 dW2 dW2 .2 Reversing the process Suppose we are given that where W1 and W2 are Brownian motions with correlation coefficient .236 23. 2 2: p 2 =) dB2 = dW2 . 2 dW1 dW2 + 2dW2 dW2 1 = 1.

=p 1 1. . 2( dW1 dW1 ) dt . Furthermore. Recognizing a Brownian Motion so both B1 and B2 are Brownian motions. 237 dB1 dB2 = p 1 1. dt) = 0: We can now apply an Extension of Levy’s Theorem that says that Brownian motions with zero cross-variation are independent. 2 (dW1 dW2 .CHAPTER 23. to conclude that B 1 B2 are independent Brownians.

238 .

We want to find 1 and 2 and define e dB1 = 1 dt + dB1 e dB2 = 239 2 dt + dB2 . q 2 2 dB2 (t) > 0 2 > 0 . where 0 < S (0) < K 0 < Y (0) < L Y (T ) = 0maxT Y (t): t Remark 24. which in this case means the discounted Y and S processes. The option pays off: (S (T ) . The risk-neutral measure must make the discounted value of every traded asset be a martingale.Chapter 24 An outside barrier option Barrier process: dY (t) = dt + Y (t) Stock process: 1 dB1 (t): dS (t) = dt + S (t) where 1 2 dB1 (t) + 1 . and B1 and B2 are independent Brownian motions on some ( F P). K )+ 1 Y (T )<L f g at time T .1 The option payoff depends on both the Y and S processes.1 < < 1. we will need the money market and two other assets. which we take to be Y and S . In order to hedge it.

1 ( 1 + 2 )T 2 Z f IP (A) = Z (T ) dIP 8A 2 F : o f e e P Under I . 2 B2 (T ) .r 1 2 1: 2 We shall see that the formulas for 1 and 2 do not matter.240 so that dY = r dt + dB 1 e1 Y = r dt + 1 1 dt + 1 dB1 dS = r dt + dB + q1 . Y S The value of the option at time zero is A f IP is the unique dY dS = YS 1 2 dt f v(0 S (0) Y (0)) = IE e h .r. rT (S (T ) . f P Remark 24. 2 e1 S = r dt + 2 1 dt + 1 .1) 1. S has volatility 2 and i. 2 2 dB2 : =r+ =r+ 1 1 2 1+ q 2 2 dB2 e We must have q (0. What matters is that (0. K )+ 1 f Y (T )<L i g : We need to work out a density which permits us to compute the right-hand side.. B1 and B2 are independent Brownian motions (Girsanov’s Theorem). 2= p 1. Y has volatility 1.1) and (0. 2 2 2: (0.2) We solve to get .e. 2 n 1 = . the correlation between dY and dS is . This implies the existence and uniqueness of the risk-neutral measure. 1 B1 (T ) .2) uniquely determine 1 and 2 . risk-neutral measure.2 Under both IP and I . . 2 2 2 dt q + 2 dB1 + 1 . We define 2 2 Z (T ) = exp .

2 dB 2 2 dB2 e b Y (t) = Y (0) expf 1B (t)g c Y (T ) = Y (0) expf 1M (T )g: The stock process. dS = r dt + S so 1 e e S (T ) = S (0) expfrT + 2B1 (T ) . 2) 2 T g 2 2 2 q e e = S (0) expfrT . is f b IP fB (T ) 2 d^ M (T ) 2 dmg b c ^ ^) ( (2m . An outside barrier option Recall that the barrier process is 241 dY = r dt + Y so e 1 dB1 n o 1 e Y (t) = Y (0) exp rt + 1B1 (t) . 2 2t : 1 Set b = r= 1 . 2 2 2B2 (T )g e . 1 2 2 T + 1 . b exp . 2 2 B2 (T )g 2 2 From the above paragraph we have q e b B1 (T ) = .CHAPTER 24. 1 2 T + 2 B1 (T ) + 1 . 2 T db dm ^ T 2 T m > 0 ^ < m: ^ b ^ q e1 + 1 . 2 (1 . 2 2 q bT + 1 . 2 2B2 (T ) . appearing in Chapter 20. ^2T b + b . bT + B (T ) so b S (T ) = S (0) expfrT + 2B (T ) . 1 2 T . ^)2 b^ 1 b2 ) ^ m ^ = 2(2p . 1=2 b e B (t) = bt + B1 (t) c b M (T ) = max B (t): 0 t T Then The joint density of b c B (T ) and M (T ).

q 2 2 2 (1. We would have had only one equation (see Eqs (0. q 2 2~ b . 2 2 b)T + b 2 B (T ) + 1 .1 . then we would not have tried to set its mean return equal to r. ~2 d~ ~ 2 IR: e IP b 2T 2 T e e b c Furthermore. Indeed. any option whose payoff depends on Y cannot be hedged when we are allowed to trade only in the stock. It also depends on 1 2 r K and L. the pair of random variables (B (T ) M (T )) is independent of B2 (T ) because B1 and e2 are independent under IP .3 If we had not regarded Y as a traded asset. 21 : Remark 24. The parameter b appearing in the answer is b = r1 .K + T 2 T :d~ d^ dm: b b ^ The answer depends on T S (0) and Y (0).1). rT (S (T ) . 2 2 2 b)T + b 2^ + 1 . The nonuniqueness of the solution alerts us that some options cannot be hedged. The density of B2 (T ) is f e f e IP fB2 (T ) 2 d~ B(T ) 2 d^ M (T ) 2 dm g = IP fB2 (T ) 2 d~g:IP fB(T ) 2 d^ M (T ) 2 dmg b b b c ^ b f b b c ^ The option value at time zero is v (0 S (0) Y (0)) =e . Therefore.K + :1 Y (0)exp 1 M (T )]<L b f g b We know the joint density of ( B (T ) ( ) b b b ffB2(T ) 2 d~g = p 1 exp .2)) = r+ 2 1+ 1. ~2 :p 2T 2 T ( ) 2(2p . h . K )+ 1 f Y (T )<L i g 1 f = e rT IE S (0) exp (r . q 2 2 B2 (T ) e . . the joint density of the random vector (B2(T ) B (T ) M (T )) f e b c B is c e M (T )). ^)2 + b^ . It does not depend on 1 nor 2 . rT 1 1 log Y L Zm Z (0) ^ 0 Z 1 ( ) b 1 exp . (2m . ^) exp . 1 b2 T m b ^ ^ b : b 2 2T .1 Computing the option value f v (0 S (0) Y (0)) = IE e . 2 2 .(0.242 24.1 S (0) exp (r .1) to determine 1 and 2 . 1 2.

2 2Svx dB2 + 1Y vy dB1 e e .r f so with dW 2 = dt + dW we have dS = r dt + S and we are on our way. h i 1 = e rt . S (0) and Y (0) by T . is superfluous. q = . and dS = dt + dW: 2 S Now we have only Brownian motion. 243 S . We compute P d e rtv(t S (t) Y (t)) . namely. K )+ 1 Y (T )<L f g we obtain a formula for v (t x y ) = e . e rt v (t S (t) Y (t)) .2 The PDE for the outside barrier option Returning to the case of the option with payoff (S (T ) . Using the Markov property. x and y respectively in the formula for v (0 S (0) Y (0)). there will be only one . then Y 1. h (S (T ) . Returning to the dS = dt + S we should set 2 dB1 + q 2 2 dB2 dW = dB1 + 1 .rv + vt + rSvx + rY vy + 2 2S 2 vxx + 2 + e 2 Svx dB1 + q 2 2 1 2 SY vxy + 1 1 Y vyy 2 dt 1.CHAPTER 24. An outside barrier option If we have an option whose payoff depends only on original equation for S . 2 f dW 24. t. 2dB2 so W is a Brownian motion under IP (Levy’s theorem). K )+ 1 maxt f u T Y (u) < Lg i by replacing T . r(T t)I t x y f E . . Now start at time 0 at S (0) and Y (0). we can show that the stochastic process f is a martingale under I .

x >= 0 x v(t. K )+ x 0 0 y<L and the boundary conditions are v(t 0 0) = 0 0 t T v (t x L) = 0 0 t T x 0: . rv + vt + rxvx + ryvy + 1 2x2vxx 2 2 + 2 2 1 2 xyvxy + 1 1 y vyy 2 =0 0 t<T x 0 0 y L: The terminal condition is v(T x y) = (x . 0.1: Boundary conditions for barrier option.244 y L v(t. L) = 0. Setting the dt term equal to 0. 0) = 0 Figure 24. Note that t 2 0 T ] is fixed. we obtain the PDE . x.

h . Note where vx that e e = vx (t S (t) Y (t)). An outside barrier option 245 . r(T t)(0 . vy = vy (t S (t) Y (t)).rY (t) dt + dY (t)] e 1Y (t) dB1 (t): . 24. i i q 2 2 Svx dB2 + 1 Y vy dB1 e e are functions of t.CHAPTER 24. The boundary condition is v (t 0 L) = 0 v(t 0 0) = 0 v (T 0 y) = (0 . The boundary conditions are the terminal condition is x=0 y=0 . K )+ = 0 y 0: On the x = 0 boundary. the option value is v (t 0 y ) = 0 0 y L: v(t 0 0) = e . d e rtv (t S (t) Y (t)) = vxd e rtS ] + vy d e rtY ]: . . The value X (t) of the portfolio has the differential dX = 2dS + 1 dY +rX. rt d e rt Y (t) = e =e . the barrier is irrelevant. 2 S . h i Let 2 (t) denote the number of shares of stock held at time t.rv + vt + ryvy + 1 2y2vyy = 0 2 1 This is the usual Black-Scholes formula in y . the terminal condition is v (T x 0) = (x .rS (t) dt + dS (t)] q e 2S (t) dB1 (t) + 1 . K )+ = 0 . K )+ x 0: On the y = 0 boundary. and B1 B2 S Y d e rt S (t) = e . . e 2Svx dB1 + 1 . 2 2S (t) dB2 (t) e : rt . we have the equation d e rtv (t S (t) Y (t)) . and the option value is given by the usual Black-Scholes formula for a European call.rv + vt + rxvx + 1 2x2vxx = 0 2 2 This is the usual Black-Scholes formula in x. and let 1(t) denote the number of “shares” of the barrier process Y . . .3 The hedge After setting the dt term to 0. h = e rt . 1Y ] dt: . h i = e rt . rt Therefore.

2 (t)d e .246 This is equivalent to d e rtX (t)] = . rtS (t)] + 1(t)d e . we must have X (0) = v (0 S (0) Y (0)) and 2 (t) = vx (t S (t) Y (t)) 1 (t) = vy (t S (t) Y (t)): . rt Y (t)]: To get X (t) = v (t S (t) Y (t)) for all t.

let x > 0 be given.2 First passage times for Brownian motion: first method (Based on the reflection principle) Let B be a Brownian motion under IP . We define vL L = minft 0 S (t) (x) = IEe r L (K . S ( = ( . and define = minft 0 B (t) = xg: is called the first passage time to x. We need to know the distribution of L . 25. Suppose we exercise the first time the stock price is L or lower. rL The plan is to comute vL (x) and then maximize over L to find the optimal exercise price. 25. K.x (K . 247 . if x > L: . L)IEe L )) Lg + if x L.1 Preview of perpetual American put dS = rS dt + S dB Intrinsic value at time t : (K . We compute the distribution of .Chapter 25 American Options This and the following chapters form part of the course Stochastic Differential Equations for Finance II. S (t))+ : Let L 2 0 K ] be given.

1: Intrinsic value of perpetual American put Define M (t) = 0maxt B (u): u From the first section of Chapter 20 we have 2 m IP fM (t) 2 dm B(t) 2 dbg = 2(2p . b) dm db t t 2 t Therefore. (2m . b) exp . b) ( (2m .1 1 x We make the change of variable z = mt in the integral to get p 2 p2 exp . b)2 b=m dm p = 2t 2 t x b= ( 2) Z p 2 exp . 2t t 2 t x ) ( Z 2 exp . (2m2. ( ) m > 0 b < m: Z Z m 2(2m . b)2 ) p IP fM (t) xg = db dm exp . mt dm: = 2 2 t 1 .248 K Intrinsic value K Stock price x Figure 25.1 1 . z2 dz: = x= t 2 1 p Z ( ) Now t()M (t) x .

CHAPTER 25. p2 exp . p 2 > 0: (See Homework) Reference: Karatzas and Shreve. xt : @t p dt 2 t 2 ( 2) = px exp . pp 196–197. = Z 1 =e x . B (t) . Brownian motion and Stochastic Calculus.3 Drift adjustment Reference: Karatzas/Shreve. More specifically. 25. define e B (t) = t + B (t) Z (t) = expf. Brownian Motion and Stochastic Calculus. 0 e t IP f 2 dtg . z 2 dz dt p = @t 2 x= t 2 ( 2) @ x = . B 2 e ~ = minft 0 B (t) = xg: A Define We fix a finite time T and change the probability measure “only up to T ”. so f IP f~ 2 dtg = IP f 2 dtg 2 = px exp . with T fixed. define Z f e Under I . x dt 0 < t T: 2t t 2 t ( ) . American Options so 249 @ IP f 2 dtg = @t IP f tg dt @ = @t IP fM (t) xg dt " Z ( ) # @ 2 exp . For 0 t < 1. pp 95-96. 1 2 tg 2 e (t) + 1 2tg = expf. the process B (t) P f IP (A) = Z (T ) dP A 2 F (T ): 0 t T . x dt: 2t t 2 t 1 p We also have the Laplace transform formula IEe . is a (nondrifted) Brownian motion.

2 2 sgIP f ~ 2 dsg 0 ( Zt x 2) 1 2s . (x . this must in fact be the correct formula for all t > 0. 2s = s 2 s f g f g f g 1 2 e ~ t expf B (T ) .4 Drift-adjusted Laplace transform Recall the Laplace transform formula for = minft 0 B (t) = xg for nondrifted Brownian motion: IEe For . s)2 ds: p exp .250 For 0 < t T we have IP f~ tg = IE 1 h h f f = IE 1 ~ t i g g f f = IE 1 i fh e = IE 1 ~ t expf B (~ ^ t) . 2 T g 1 f f e = IE 1 ~ t IE expf B (T ) . t t) dt 0 < t T: 2 t 2 t Since T is arbitrary. xt dt = e 2 t 2 t ( ) . = Z 0 1 2 px exp . 2 IP f~ 2 dtg = px exp . 1 2 (~ ^ t)g 2 i fh = IE 1 ~ t expf x . 2 2 T g F (~ ^ t) f g ~ t Z (T ) 1 i 0 Therefore. t . x 2 p > 0 x > 0: ~ = minft 0 t + B (t) = xg . x p exp x . 2 = 2s ds 0 s 2 s ) ( Zt x (x . 1 2 ~g 2 Zt 1 f = expf x . ( ) 25.

if ~(! ) = 1. then (Recall that x > 0). ~= 2 px exp . 1 2 t dt p = 2 2t 0 t 2 t ) ( Z x exp . so lim e ~(! ) = 0: . t t) dt 2 0 t 2 t ( ) Z x exp . Therefore. lim e . (x . #0 ~(! ) = 1 ~< 1 : 2 ~ = ex x 2 + p . ~(! ) = 1 > 0. we obtain If If IP f~ < 1g = ex x p 2 = ex x : . Z ( ) where in the last step we have used the formula for IEe.5 First passage times: Second method (Based on martingales) Let > 0 be given. x2 + x .( + 1 2 )t . 25. Then Y (t) = expf B(t) . . t . lim e Letting #0 and using the Monotone Convergence Theorem in the Laplace transform formula IEe . . x2 dt x p =e 2 2t 0 t 2 t = ex x 2 + 2 >0 x>0 1 1 1 p . American Options the Laplace transform is 251 IEe .CHAPTER 25. j j 0. If ~(!) < 1. 1 2tg 2 . then IP f ~ < 1g = 1: IP f ~ < 1g = e2x < 1: < 0. t . then with replaced by +1 2 2. then e ~(!) = 0 for every #0 #0 .

1 2 2 1< 1 1 where we understand e x. 1 lim expf B (t ^ ) . 1 1 2 lim expf B (t ^ ) . x 2 p = IEe >0 x>0 for the first passage time for nondrifted Brownian motion. 1 1 = IE t! expf B (t ^ ) . Since 0 expf B (t ^ ) .252 is a martingale. For those ! for which (! ) < 1. so Y (t ^ ) is also a martingale. 1 2 (t ^ )g: 2 1 = t! IE expf B (t ^ ) . . . 1 2 (t ^ )g = e x 2 : 2 t! For those ! for which (! ) = 1. 1 2 )t + B(t)g 2 8 > > < = x exp > > : 2 39 > > 6 r 6 . 25.6 Perpetual American put dS = rS dt + S dB S (0) = x S (t) = x expf(r . 1 2 (t ^ )g: lim 2 There are two possibilities. . Therefore. 2 2tg = 0: lim 2 t! 1 1 Therefore. 1 2 (t ^ )g 2 1 ex this is justified by the Bounded Convergence Theorem. 1 2 (t ^ )g t! expf x . so = 2 p to be zero if = 1. We have again derived the Laplace transform formula e . 2 2(t ^ )g lim 1 = IE e x . We have 1 = Y (0 ^ ) = IEY (t ^ ) = IE expf B (t ^ ) . t + B(t)7= : 7 6 7> 4| {z 2 } 5> . 2 Let = IEe x 2 1 2 2 =1 2 2 . . 1 = IE t! expf B (t ^ ) . 2 2 (t ^ )g: lim 1 We want to take the limit inside the expectation.

log L . L) exp . r2 + 1 . p . 2 8 <(K . L) L . American Options Intrinsic value of the put at time t: Let L 2 253 0 K ] be given. S (t)) +. t . L) L 2r= 2 are all of the form Cx. L)L2r= 2 . L) L . r2 + 1 . 2r= 2 x L: 2 0 x L . r + 1 . . 1 r 2 + r + 2=4 2 = . 1 r + =2 2 = . 1 2r + 2 = . 0 S (t) = Lg = minft 0 t + B (t) = 1 log L g x 1 log x g = minft 0 . 2r : 2 Therefore. r2 + 1 . Define for x L. r + 2 =4 2 2 2 = .x The curves (K . 1 2r + r . We want to choose the largest possible constant. 1 2r + r 2 . r2 + 1 . 2 We compute the exponent p . . x) vL (x) = : . =2 2 2 s s s s 2 = . 1 r + =2 2 = .CHAPTER 25. B (t) = L Define vL = (K . L = minft (K . . 1 log L 2r + 1 2r + 2 x : = (K . The constant is C = (K . L)IEe r L x xp = (K .x (K .2r= .

2: Value of perpetual American put value C3 x C2 x C1 x -2r/ σ 2 -2r/ σ 2 -2r/ σ 2 Stock price x Figure 25.3: Curves.L) (x/L)-2r/ σ K Stock price x Figure 25.254 value K-x K 2 (K . .

2 r2 K .1 = lim v (x): x"L 0 . 2 (K . 2r= 2 x L where 2 L = 2 rK2r : + Note that v (x) = . We solve . L)L 2r2 1 2 @L 2r 1 = L 2 .1 + 2r (K .CHAPTER 25.4): 8 < v (x) = :(K .2 r2 (K . L ) L ( . 2r ! 2 + 2r = .2 r2 2 + 2r 2r = . L) (L 0 )2r= 2 x . 1 + 2r + 2r K = 0 2 2L to get 2 L = 2 rK2r : + Since 0 < 2r < 2 + 2r we have 0 < L < K: 0 x L Solution to the perpetual American put pricing problem (see Fig. L ) 1 L 2+ 2 = . x) . American Options and 255 @C = .L 2r2 + 2r (K . 2r= 2. 2 rK2r 2rK2r + 2 + 2r . x (K . L) L 2 2r = L 2 . 1 0 x<L x>L : We have x#L lim v (x) = .12r 0 . 1 + 2r + 2r K : 2 2L . 25.

In other words.256 value K-x K -2r/ σ (K . x. L )+1 .1) (7. 1 2 2r 2 2 = 0: 0 0 00 .r . 25.2) = IE x where h e (K . If L x < 1.7 Value of the perpetual American put Set 2 L = 2 rK2r = + 1 K: + If 0 x < L .5). . then . . 25. r . 1)x 2 ] 2 1 2 (. .4: Solution to perpetual American put. = 2r 2 (7. then v(x) = (K .rv(x) + rxv (x) + 1 2x2v (x) 2 = C .rv(x) + rxv (x) + 1 2x2v (x) = .4) x < L .rK: 2 If L x < 1. x) + rx(.1) = .3) (7. . L )(L ) } x | {z . rx x 1 . . 2 = C (.rx .L* )(x/L* ) 2 L* K Stock price x Figure 25. . C r f < i 1g S (0) = x If 0 = minft 0 S (t) = L g: 00 (7. then . . . then v (x) = K . . v solves the linear complementarity problem: (See Fig. 1 2 x2 (. 1)x r . 1)] = Cx . .r(K .

1 2x2v > 0g 2 0 00 and L is the boundary between them. rxv . rt (t) S (t) dB (t): . Invest the money. 2 2x2 v 0 v (K . The value X (t) of this portfolio is governed by dX (t) = (t) dS (t) + r(X (t) . the owner of the put should not exercise (should “continue”). If the stock price is in S or at L . holding (t) shares of stock and consuming at rate C (t) at time t.8 Hedging the put Let S (0) be given. . Sell the put at time zero for v (S (0)). (t)S (t)) dt . The half-line 0 1) is divided into two regions: C = fx v(x) > (K . rxv . 1 rv . If the stock price is in C .e rt C (t) dt + e .CHAPTER 25. d(e rt X (t)) = . . x)+ 0 00 (a) (b) (c) One of the inequalities (a) or (b) is an equality. C (t) dt or equivalently.5: Linear complementarity For all x 2 IR. American Options 257 6 K@ @@ L v @@ @@ K -x Figure 25. the owner of the put should exercise (should “stop”). x 6= L . x)+ g S = fx rv . 25.

S (t) (t) = v (S (t)) = . e rtv (S (t)) is a supermartingale (see its differential above). then As long as the owner does not exercise. Explanation of property 3.rKe rt 1 S (t)<L dt + e rt S (t)v (S (t)) dB (t): . S (t))+ .3) If t is not zero. .258 The discounted value of the put satisfies d e rtv(S (t)) = e . v(S (t)) = K . 2.2) for t = 0. and then .1) Then property 3 says that Y (t) e rt v (S (t)) 0 t < 1: (8. i. S (t))+ .3). short one share of stock and hold K in the money market. we can take t to be the initial time and S (t) to be the initial stock price... rt + e rt S (t)v (S (t)) dB (t) = . 0 C (t) = rK 1 S(t)<L : f g Properties of e. S (0))+ = v(S (0)): (8:1) .1) to prove (8.rt v (S (t)): 1. e rtv (S (t)) is the smallest process with properties 1 and 2.2). . 3. adapt the argument below to prove property (8. assuming Y is a supermartingale satisfying (8.e.1): Case I: S (0) L : We have Y (0) |{z}(K .e. i. .rv(S (t)) + rS (t)v (S (t)) + 1 2S 2(t)v (S (t)) dt 2 0 00 We should set C (t) = rK 1 S (t)<L (t) = v (S (t)): f 0 g Remark 25. e rtv (S (t)) e rt (K .1 If S (t) < L . . 0 t < 1. . you can consume the interest from the money market position. Y (0) v(S (0)): (8. 0 f g h i .2) We use (8. Let Y be a supermartingale satisfying Y (t) e rt (K . Proof of (8. .1: To hedge the put when S (t) < L . 0 . 0 t < 1: (8.

Value of the American contingent claim: v (x) = sup IE x e r h(S ( )) .CHAPTER 25. . . f = v (S (0)): L < 1g 3 7 5 (by 8. 7. S{z ))+ 1 4 |( } f 1g . .1) (See eq.9 Perpetual American contingent claim Intinsic value: h(S (t)). . Theorem 10. which attains the supremum. e rtv (S (t)) is the smallest process with properties 1 and 2.10 Perpetual American call v(x) = sup IE x e r (S ( ) . American Options Case II: S (0) > L : For T 259 > 0.2) 25. we have f 1g Y (0) IEY ( ^ T ) (Stopped supermartingale is a supermartingale) h i IE Y ( ^ T )1 < : (Since Y 0) Now let T !1 to get Y (0) lim IE Y ( ^ T )1 < T! h i (Fatou’s Lemma) IE Y ( )1 < 1 f 1g h i 2 IE 6e r (K . 3. e rtv (S (t)) e rt h(S (t)) 0 < t < 1. where the supremum is over all stopping times.63 v (x) = x 8x 0: . 25. Optimal exercise rule: Any stopping time Characterization of v : 1. e rtv (S (t)) is a supermartingale. K )+ . 2.

IE x e r (S ( ) .6. It can be shown that jump.2 No matter what .11 Put with expiration T > 0. Expiration time: Value of the put: v(t x) = (value of the put at time t if S (t) = x) = sup IE xe r( t) (K . x v(x): we choose. K )+ 0 t < 1. 25. Now start with S (0) = x and define Y (t) = e rtS (t): . v(x) IE x e rt (S (t) . h i x. . Y is a supermartingale (in fact. Remark 25. . K ) h i = IE x e rt S (t) . . Therefore. Y (t) e rt(S (t) . Let t!1 to get v (x) Then: 1. e rt K = x . . S ( ))+: .e. i. K )+ < IE x e r S ( ) There is no optimal exercise time..260 Proof: For every t. . while v xx has a . Y (0) v(S (0)). . Let S (0) be given. Then | {z } T v vt vx are continuous across the boundary. 2. S (t)) +. t :stopping time See Fig. x = v(x): 25. Y is a martingale). e rt K: . . Intrinsic value: (K . K )+ h i IE x e rt (S (t) .

American Options 261 x 6 v>K. Expiration time: Value of the contingent claim: v(t x) = sup IE x e t T Then . r( t)h(S ( . x 0 x K -t Figure 25. e rtv (t S (t)) e rt (K .12 American contingent claim with expiration T > 0. . vt .rv + vt + rxvx + 1 2x2vxx = . v (T x) = K .rK 2 T 1. S (t))+ 0 t T .CHAPTER 25.rv + vt + rxvx + 1 2x2vxx = 0 2 K v (T x) = 0 x K L v=K. 25.6: Value of put with expiration e rtv (t S (t)) 0 t T is a supermartingale. . Then . 3. rv .x . either (a) or (b) is an equality.x vt = 0 vx = . )): At every point (t x) 2 0 T ] 0 1). 1 2 x2vxx 0 2 v h(x) (a) (b) (c) Characterization of v : Let S (0) be given. 2. . Intrinsic value: h(S (t)). rxvx . .1 vxx = 0 . e rtv (t S (t)) is the smallest process with properties 1 and 2.

. 2. e rtv (t S (t)) e rt h(S (t)). . e rtv (t S (t)) 0 t T is a supermartingale. 3. The optimal exercise time is = minft 0 v (t S (t)) = h(S (t))g If (!) = 1. . e rtv (t S (t)) is the smallest process with properties 1 and 2. then there is no optimal exercise time along the particular path ! . . .262 1.

g. This stock pays no dividends.e.Chapter 26 Options on dividend-paying stocks 26. and assume h(0) = 0. 0 and S (T ) (t) = exp . An Let h(x) be a convex function of x American contingent claim paying h(S (t)) if exercised at time t does not need to be exercised before expiration. )h(0) + h(x) = h(x): Let T be the time of expiration of the contingent claim. The value of this claim at time t 2 0 T ] is X (t) = (t) IE (1 ) h(S (T )) F (t) : T 263 .s. (E. Proof: For 0 1 and x 0.64 Consider the stock price process dS (t) = r(t)S (t) dt + (t)S (t) dB (t) where r and are processes and r(t) 0 0 t T a.. )0 + x) (1 . h(x) = (x . i.1 American option with convex payoff function Theorem 1. so h (t) h(S (T )): (T ) (*) Consider a European contingent claim paying h(S (T )) at time T . waiting until expiration to decide whether to exercise entails no loss of value. Z T r(u) du (T ) t (t) S (T ) (T ) ( ) 1 0. 0. For 0 t T . K )+ ). we have h( x) = h((1 ..

. . . (x h(x)) .. .. .. . . .... ..... ........ . ... .. .. i...... . ...... ... ...... .... .. ..... ... .........r.. X (t) = 1 IE (t) h(S (T )) F (t) (t) (t) (T ) (t) 1 (t) IE h (T ) S (T ) F (t) (by (*)) 1 h (t) IE S (T ) F (t) (Jensen’s inequality) (t) (T ) = 1 h (t) S (t) ( S is a martingale) (t) (t) 1 = (t) h(S (t)): This shows that the value X (t) of the European contingent claim dominates the intrinsic value h(S (t)) of the American claim... ................... . . ..... the American claim should not be exercised prior to expiration............. ............ except in degenerate cases... . ............ . ..... .. ........ .... .. .... . .. . ..... .. . h .. . ..... ...... h( x) .. .264 . .. ....... .. ... .. ...... . .. . .... .. ... .............. .... ..... ..... ... let be a “dividend coefficient” satisfying 0 < < 1: . .. .... .... ............ .2 Dividend paying stock Let r and be constant...... . ... . .. .r...... ... . .............r.. . .. ........ .......... .... ...... 6 x - Figure 26.1: Convex payoff function Therefore.. 26.. ............ h(x) .......... .... .. ... the inequality X (t) h(S (t)) 0 t T is strict..... In fact.........e... ........

2 t1 < t T: Consider an American call on this stock. t x)) . it is not optimal to exercise. )S (t1) expf(r . )x : t t1 . The stock 1 2 0 t t1 S (t) = S (0) expf(r .CHAPTER 26. the value of the call is w(t1 S (t1)) where Theorem 2. and let t1 price is given by 265 2 (0 T ) be the time of dividend payment. K )+ v (t1 (1 .t v (t1 (1 . . K )+ v(t1 (1 . the value of the American call is w(t S (t)). r(T t) N (d . immediately after payment of the dividend.65 For 0 w(t1 x) = max (x . so the value of the call is given by the usual Black-Scholes formula ( v(t x) = xN (d+(T .64. )S (t1)): 2=2) At time t1 . the value of the call is x d (T . )x) x 0 w(t 0) = 0 0 t T: (t) = and boundary condition The hedging portfolio is ( wx(t S (t)) vx (t S (t)) 0 t t1 t1 < t T: Proof: We only need to show that an American contingent claim with payoff w(t 1 S (t1)) at time t1 need not be exercised before time t1. t)(r T . . 1 S (t1)) : i This function satisfies the usual Black-Scholes equation . t1 ) + (B (t) . Options on dividend paying stocks Let T > 0 be an expiration time. Ke where . immediately before payment of the dividend.rw + wt + rxwx + 1 2x2wxx = 0 0 t t1 x 0 2 (where w = w(t x)) with terminal condition w(t1 x) = max (x . B (t1 ))g (1 . According to Theorem 1. r(t1 t) w(t . it suffices to prove 1. At times t 2 (t 1 T ). (T . where w(t x) = IE t x e h . t x)) t1 < t T : At time t1 . t x) = p 1 log K + (T . w(t1 0) = 0. 2 )t +1 B(t)g 2 )(t .

The proof of the convexity of v (t1 (1 .266 2. Obviously. . K before the dividend payment at time t1 . (t1 +) = 1 (t1) = (t1) + 1. ) (t1+) where (t1 +) = lim (t) t#t1 is the number of shares of stock held by the hedge immediately after payment of the dividend.3 Hedging at time t1 Case I: v (t1 (1 . Indeed. We have Let x = S (t1). 1. )0) = 0: To prove that w(t1 x) is convex in x. and the maximum of two convex functions is convex. K )+ . )x) is convex is x. )vx (t1 (1 . (x . )x) < (x . K ). K to v (t 1 (1 . )x) (t1) = wx (t1 x) = (1 . 0) = 0. )S (t ) 1 (t1) dividends received price per share when dividend is reinvested = # of shares held when dividend is paid + Case II: v (t1 (1 . )x) = (1 . 26. K )+ v(t1 (1 . its value drops from x . (t1 )S (t1) = (t1 ) + (1 . The option need not be exercised at time t 1 (should not be exercised if the inequality is strict). we need to show that v (t 1 (1 . )x) in x is left as a homework problem. we have immediately that Since v (t1 w(t1 0) = max (0 . w(t1 x) is convex in x. and the seller of the option can pocket the difference and continue the hedge. If the option is not exercised. The owner of the option should exercise before the dividend payment at time t 1 and receive (x . The post-dividend position can be achieved by reinvesting in stock the dividends received on the stock held in the hedge. )x) (x . The hedge has been constructed so the seller of the option has x . K )+ is convex in x. K )+ . )x). w(t1 x) = v (t1 (1 .

Con- dS (t) = r(t)S (t) dt + (t)S (t) dW (t): Here. pays 1 at time T and nothing before time T . ZT t r(u) du F (t) : ) # Given B (t T ) dollars at time t. forward contracts and futures Let fW (t) F (t) 0 t T g be a Brownian motion (Wiener process) on some ( sider an asset. Define the accumulation factor (t) = exp Zt 0 r(u) du : A zero-coupon bond. which we call IP . its value at time t 2 0 T ] is B(t T ) = (t) IE (1T ) F (t) t = IE ((T)) F (t) " ( = IE exp .Chapter 27 Bonds. whose price satisfies F P). and we have already switched to the risk-neutral measure. r and are adapted processes. maturing at time T . which we call a stock. one can construct a portfolio of investment in the stock and money 267 . Assume that every martingale under IP can be represented as an integral with respect to W . According to the risk-neutral pricing formula.

then is not zero. If.e. and the two which are most convenient can depend on the instrument being hedged. then we will have X (T ) = B(T T ) = 1: If r(t) is nonrandom for all t. = 0. B(t T ) = (t) IE (1 ) F (t) | T {z } martingale 1 + Z t (u) dW (u) = (t) IE (T ) Zt 0 = (t) B (0 T ) + (u) dW (u) 0 dB (t T ) = r(t) (t) B (0 T ) + (u) dW (u) dt + (t) (t) dW (t) 0 = r(t)B (t T ) dt + (t) (t) dW (t): The value of a portfolio satisfies Zt dX (t) = (t) dS (t) + r(t) X (t) .268 market so that the portfolio value at time T is 1 almost surely. Then given above is zero. the money market. (t)S (t)]dt = r(t)X (t) dt + (t) (t)S (t) dW (t): (*) We set (t) (t) = (t)S(t) : (t) If. then ( ZT ) B (t T ) = exp .. Any two of them are sufficient for hedging. r(u) du t dB (t T ) = r(t)B(t T ) dt i. for some process . Indeed. you are given B (t invest only in the money market. X (t) = B (t T ) and we use the portfolio (u) t u T . at any time t. then at time T you will have T ) dollars and you always B(t T ) exp (Z T t r(u) du = 1: ) If r(t) is random for all t. at time t. One generally has three different instruments: the stock. and the zero coupon bond. .

1 (Value vs. F (t) is the price agreed upon at time t which will be paid for the stock at time T . i.1 Forward contracts We continue with the set-up for zero-coupon bonds. The T -forward price of the stock at time t 2 0 T ] is the F (t)-measurable price. its value at time t 2 0 T ] is V (t) = (t) IE (1 ) (S (T ) . In fact.CHAPTER 27. F (t) B (t T ): (t) (t) This implies that ) F (t) = BSt(tT ) : ( Remark 27. At later times. Forward price) The T -forward price F (t) is not the value at time t of the forward contract. it does not. F (0)) F (t) T t (T ) = (t) IE S(T ) F (t) .2 Hedging a forward contract Enter a forward contract at time 0. We solve for F (t): IE (1 ) (S (T ) . F (0)B (t T ) = S (t) . This generates income . At time 0. (0 T At time zero. F (0)B (t T ): F (0)B (0 T ) = BS (0) ) B (0 T ) = S (0): (0 T This suggests the following hedge of a short position in the forward contract. F (t) IE (t) F (t) (T ) (t) (T ) = S (t) . agreed upon at time t. for purchase of a share of stock at time T . agree to pay F (0) = BS (0) ) for a share of stock at time T . F (0) IE ((T)) F (t) (t) = (t) S(t) .. In other words.e. F (t)) F (t) (T ) = IE S (T ) F (t) . The value of the contract at time t is zero. 27. forward contracts and futures 269 27. short F (0) T -maturity zero-coupon bonds. however. chosen so the forward contract has value zero at time t. this contract has value 0. Bonds. F (t)) F (t) = 0 0 t T: T 0 = IE 1 (S (T ) .

either party can close out his/her position at any time. Maintain this position until time T . ) The mechanism for receiving and making these payments is the margin account held by the broker. at times tk+1 tk+2 : : : tn . you receive a payment (tk+1 ) . The value at time t = t0 of this sequence is 2n 1 X (t) IE 4 .3 Future contracts Future contracts are designed to remove the risk of default inherent in forward contracts. F (0)B (T T ) = S (T ) . when the portfolio is worth S (T ) . F (0): Deliver the share of stock and receive payment F (0). Let us first consider the situation with discrete trading dates 0 = t0 < t1 < : : : < tn = T: On each tj tj +1 ). but the implementation of this hedge would require a term-structure model. you make the payment . (tk ). = tn . Thus. Through the device of marking to market. By time T is F (tk )-measurable. (If the price has fallen. when the future price is (tk+1 ). the value of the future contract is maintained at zero at all times. A short position in the forward could also be hedged using the stock and money market.( (tk+1 ) . (tk+1 ) : : : (tn ) . so (tk+1 ) = exp Z tk+1 8 0k 9 <X = = exp : r(tj )(tj +1 . (tj )) F (t) : Because it costs nothing to enter the future contract at time t. This portfolio requires no initial investment. you have received the sequence of payments (tk+1 ) . r is constant. when the future price is (t k ). (tn 1 ) . tj ) j =0 r(u) du Enter a future contract at time tk . j =k 3 1 5 (tj +1 ) ( (tj +1 ) . 27. taking the long position. (tk )). this expression must be zero almost surely.270 Buy one share of stock. . At time tk+1 . (tk ) (tk+2 ) .

.s. Bonds. Z t 1 d (u) (u) 0 (u) 0 (u) = 0: M (t) = IE On the other hand. if (b) holds. then IE "Z T t # 1 d (u) F (t) = IE Z t 1 d (u) F (t) . If is a martingale.66 The unique process satisfying (a) and (b) is (t) = IE S (T ) F (t) 0 t T: Proof: We first show that (b) holds if and only if R t 1 d (u) is also a martingale. and # (a) (b) 0 t T: Theorem 3. then the martingale "Z T 0 1 d (u) F (t) (u) # satisfies "Z T # Zt 1 1 d (u) F (t) M (t) = d (u) + IE 0 (u) t (u) Zt 1 = 0 (u) d (u) 0 t T: this implies 1 dM (t) = (t) d (t) d (t) = (t) dM (t) . as it should be when approximating a stochastic integral.1 The T -future price of the stock is any F (t)-adapted stochastic process f (t) 0 t T g satisfying IE "Z T t 1 d (u) F (t) = 0 (u) (T ) = S (T ) a. so 0 (u) is a martingale.CHAPTER 27. forward contracts and futures The continuous-time version of this condition is 271 (t) IE "Z T t 1 d (u) F (t) = 0 (u) # 0 t T: Note that (tj +1 ) appearing in the discrete-time version is F (t j )-measurable. Definition 27.

entered at time 0. the buyer agrees to pay F (0) for an asset valued at S (T ). (0): ( (0) . entered at time 0.5 Forward-future spread Future price: Forward price: (t) = IE S (T ) F (t) . ) F (t) = BSt(tT ) = ( (t)IE S (t) 1 (T ) F (t) : Forward-future spread: h 1 (0) . Now define (t) = IE S (T ) F (t) 0 t T: Clearly (a) is satisfied. he has paid a total of for an asset valued at S (T ). Indeed. then he pays S (T ) at time T for an asset valued at S (T ). With a future contract. the buyer receives a cash flow (which may at times be negative) between times 0 and T . The cash flow received between times 0 and T sums to ZT 0 d (u) = (T ) . IE S (T ) (T ) (T ) : If (1 ) and S (T ) are uncorrelated. S (0) i = IE 1 IE (T ) 1 (T ) IE 1 IE (S (T )) . The only payment is at time T . 27.4 Cash flow from a future contract With a forward contract. is a martingale. F (0) = IE S (T )] . if the future contract holder takes delivery at time T . S (T )) + S (T ) = (0) Thus. so (b) is also satisfied. If he still holds the contract at time T . T (0) = F (0): . is the only martingale satisfying (a).272 and so is a martingale (its differential has no dt term). (0) = S (T ) . this 27. By the tower property.

CHAPTER 27.6 Backwardation and contango Suppose Define = . and P f dS (t) = rS (t) dt + S (t) dW (t): We have Because (1 ) T f = e rT is nonrandom. (t) = ert 1 S (t) = S (0) expf( . The owner of the future tends to receive income when the stock price rises. r dS (t) = S (t) dt + S (t) dW (t): f W (t) = t + W (t). He withdraws from the money market to do this just as the interest rate rises. 1 2 )t + W (t)g 2 f (t) = IE S (T ) F (t)] = F (t) ) = BSt(tT ) = er(T t)S (t): ( . W (T ) . S (T ) and (1T ) are uncorrelated under IP . In short. 27. 2 2)t + W (t)g f = S (0) expf(r . If the stock price falls. Bonds. Therefore. forward contracts and futures If (1 ) and S (T ) are positively correlated. 1 2T + W (T ) 2 T S (0): =e h n oi . . The buyer of the future is T compensated by a reduction of the future price below the forward price. 1 2 T g 2 Z f IP (A) = Z (T ) dIP 8A 2 F (T ): A f f Then W is a Brownian motion under I . The expected future spot price of the stock under IP is IES (T ) = S (0)e T IE exp . but invests it at a declining interest rate. the long position in the future is hurt by positive correlation between (1 ) and S (T ). then T 273 (0) F (0): This is the case that a rise in stock price tends to occur with a fall in the interest rate. the owner usually must make payments on the future contract. Z (T ) = expf.

(0) = erT S (0): < r.274 The future price at time 0 is If > r. then (0) < IES (T ): This situation is called normal backwardation (see Hull). If then (0) > IES (T ). . This is called contango.

and fF (t) 0 t T g is the filtration generated by W .e. Suppose we are given an adapted interest rate process fr(t) lation factor Z 0 t T g. We define the accumu- (t) = exp t 0 r(u) du 0 t T: In a term-structure model. We assume these bonds are default-free and pay $1 at maturity. Theorem 0. let B(t T ) = price at time t of the zero-coupon bond paying $1 at time T . such that for each T 2 (0 T ]. we take the zero-coupon bonds (“zeroes”) of various maturities to be the primitive assets.Chapter 28 Term-structure models Throughout this discussion. Therefore 1 1 d B(t(tT ) = B(t T ) d (t) + (t) dB(t T ) ) = (t T ) . P Remark 28. r(t)] B (t(tT ) dt + (t T ) B (t tT ) dW (t) ) () 275 . equivalent to IP )..67 (Fundamental Theorem of Asset Pricing) A term structure model is free of arbif trage if and only if there is a probability measure I on (a risk-neutral measure) with the same P probability-zero sets as IP (i. For 0 t T T . fW (t) 0 t T g is a Brownian motion on some probability space ( F P).1 We shall always have dB (t T ) = (t T )B (t T ) dt + (t T )B(t T ) dW (t) 0 t T for some functions (t T ) and (t T ). the process B(t T ) (t) 0 t T f is a martingale under I .

we should change to a measure I under which the mean rate of return is r(t). HullWhite). If we want to have n-factors. IP is a risk-neutral measure 28. Cox-Ingersoll-Ross.1 Computing arbitrage-free bond prices: first method Begin with a stochastic differential equation (SDE) dX (t) = a(t X (t)) dt + b(t X (t)) dW (t): The solution X (t) is the factor. we take r(t) to be X (t) (e.. T ) has mean rate of return r(t) under IP . then the model admits an arbitrage by trading in zero-coupon bonds.2 Some interest-rate dependent assets Coupon-paying bond: Payments P1 P2 : : : Pn at times T1 T2 : : : Tn. 28. we let W be an n-dimensional Brownian motion and let X be an n-dimensional process. is the interest rate r(t).276 so IP is a risk-neutral measure if and only if (t T ). Option expires at time T1 < T. (t) IE 1 (B (T T ) . In the usual one-factor models. If the mean rate of return of B (t T ) under IP is not r(t) at each time t and for f each maturity T . K )+ F (t) 1 (T1) 0 t T1: . the mean rate of return of B (t T ) under IP . we define the zero-coupon bond B (t T ) = (t) IE (1T ) F (t) " ( = IE exp . Price at time t is X f k:t<Tk Pk B (t Tk ): g Call option on a zero-coupon bond: Bond matures at time Price at time t is T.g. We let the interest rate r(t) be a function of X (t). Since B (t and there is no arbitrage. ZT t r(u) du F (t) # 0 t T T: We showed in Chapter 27 that dB (t T ) = r(t)B(t T ) dt + (t) (t) dW (t) for some process . If P such a measure does not exist. Now that we have an interest rate process prices to be fr(t) 0 t ) T g.

at time t buy a T -maturity zero and short BBt(t T ) ) (T + )-maturity zeroes.CHAPTER 28. @T log B(t T ): The forward rate is #0 (4. B(t (T T ) ) B(t T + ) = 0: + At time T . R(t T T + ) = .1 (Term-structure model) Any mathematical model which determines. Y (t T ) = . you receive $1 from the T -maturity zero.4 Forward rate agreement Let 0 t T < T + T be given. At time T + . t T T .1) . T 1 t log B(t T ): . To synthesize a forward-rate agreement to do this. The ( T+ effective interest rate on the dollar you receive at time T is R(t T T + ) given by B (t T ) = expf R(t T T + )g B (t T + ) or equivalently.2 (Yield to maturity) For 0 F (t)-measurable random-variable satisfying Y (t T ) is the B(t T ) exp f(T . you pay $ BBt(t T ) ) . at an interest rate agreed upon at time t.3 Terminology Definition 28. at least theoretically. the stochastic processes B(t T ) 0 t T for all T 2 (0 T ]. log B (t T ) : @ f (t T ) = lim R(t T T + ) = . t)Y (t T )g = 1 or equivalently. B (t T ) 0 t T Y (t T ) 0 t T T T: Determining is equivalent to determining 28. Suppose you want to borrow $1 at time T with repayment (plus interest) at time T + . Term-structure models 277 28. ( T+ The value of this portfolio at time t is Bt B(t T ) . the yield to maturity Definition 28. log B (t T + ) .

Integrating the above equation. If you invest $ B (t at time t and receive interest rate f (t u) at each time u between t and T . agreed upon at time t. Z T r(u) du F (t) @T t " @ @T B(t T ) T =t = . log B (t u) = .5 Recovering the interest r (t) from the forward rate @ @T B(t T ) T =t = IE .r(t): On the other hand. f (t u) du : t You can agree at time t to receive interest rate f (t u) at each time u 2 t T ].278 This is the instantaneous interest rate.f (t t): Conclusion: r(t) = f (t t). Z T f (t u) du @T t . ( ZT ) B (t T ) = exp . f (t u) du t ( ) @ B (t T ) = . ( ZT ) # B (t T ) = IE exp . this will grow to T) B(t T ) exp at time T . ZT @ @u log B(t u) du t u=T u=t = . r(u) du F (t) t " ( ) # @ B (t T ) = IE . log B (t T ) so ( ZT ) B(t T ) = exp .r(T ) exp . (Z T t f (t u) du = 1 ) 28. we obtain ZT t f (t u) du = . for money borrowed at time T .r(t) F (t) = .f (t T ) exp .

Recall that df (t T ) = (t T ) dt + (t T ) dW (t): ( ZT ) B(t T ) = exp . dW ) dB (t T ) = dg . 0 ZT t 00 t ZT f (t u) du f (t u) du (r dt . (t T )B(t T ) dW (t): h . df (t u) du t t ZT = r(t) dt . (t T ) + 1 ( (t T ))2 dt 2 . Term-structure models 279 28. t t {z } | {z } | = r(t) dt . (t T ) dW (t): (t T ) (t T ) Let Then g(x) = ex g (x) = ex g (x) = ex: 0 00 B (t T ) = g . and ZT t f (t u) du ! ! dt . =g . (t T ) dt .6 Computing arbitrage-free bond prices: Heath-Jarrow-Morton method For each T 2 (0 T ]. (t u) du dt . let the forward rate be given by Zt Zt f (t T ) = f (0 T ) + 0 (u T ) du + Here f (u T ) 0 u T g and f (u T ) 0 u T g are adapted processes.CHAPTER 28. f (t u) du = f (t t) dt . ! + 1g 2 = B (t T ) r(t) . ZT t f (t u) du ( )2 dt ! i . f (t u) du : t Now ( ZT ) ZT d . 0 (u T ) dW (u) 0 t T: In other words. (t u) dt + (t u) dW (t)] du t "Z T # "Z T # (t u) du dW (t) = r(t) dt .

be a deterministic function. Let f (t) 0 t T g be an adapted process. let (u T ) 0 u T (u T ) > 0 for all u and T . and assume f (0 T ) 0 t T .3) (7.t. be adapted processes.1) imply (7.. T yields the equivalent condition (t T ) = (t T ) (t T ) + (t T ) (t) 0 t T T : Theorem 7. but there might still be a riskneutral measure.r. we must have P (t T ) = 1 ( (t T ))2 + (t T ) (t) 0 t T T : 2 Differentiation w. we obtain (t T ) = (t T ) ZT t (t u) du (7. A Z (T ) dIP 8A 2 F (T ): Zt 2 1 2 0 (u) du i dB (t T ) = B(t T ) r(t) .1) ZT t (t u) du = 1 2 ZT t (t u) du !2 0 t T 0 t T Differentiating this w. Suppose (7. and define f W (t) = f IP (A) = Then Zt 0 (u) du + W (t) Z (t) = exp .7 Checking for absence of arbitrage IP is a risk-neutral measure if and only if (t T ) = 1 ( (t T ))2 2 i.2) also implies (7. 0 t T T T: T: (7.68 (Heath-Jarrow-Morton) For each T 2 (0 (u T ) 0 u T .280 28. This will be a homework problem. (h T )B(t T ) dW (t) t i = B (t T ) r(t) .r.e. (u T ) dW (u): f (t T ) = f (0 T ) + Zt 0 (u T ) du + Zt 0 .2). and define h (7.2) Not only does (7. Then IP is not a risk-neutral measure.4) and Let T ].t.1) does not hold. (t T ) + 1 ( (t T ))2 + (t T ) (t) dt 2 f . Zt 0 Z (u) dW (u) . (t T )B(t T ) dW (t) 0 t T: f In order for B (t T ) to have mean rate of return r(t) under I .1). (7. T . (t T ) + 1 ( (t T ))2 dt 2 .

we may plug (7.2 Under IP . we obtain for every maturity T : f dB(t T ) = r(t)B(t T ) dt . (t T )B(t T ) dW (t): Because (7.3). (t T ) + 2 ( (t T ))2 : (t T ) The no-arbitrage condition is that this market price of risk at time t does not depend on the maturity T of the bond. the zero-coupon bond with maturity T has mean rate of return r(t) . satisfying (7. (t T ) + 2 ( (t T ))2 and when normalized by the volatility. so we can solve (7. Remark 28. or equivalently.4).3) is satisfied.3) for . The excess mean rate of return. this becomes the market price of risk 1 .4) into the stochastic differential equation for f (t T ) to obtain. Term-structure models 281 t T T is a family of forward rate processes for a term-structure model Then f (t T ) 0 without arbitrage if and only if there is an adapted process (t) 0 t T . (The remainder of this chapter was taught Mar 21) " # Suppose the market price of risk does not depend on the maturity T . Plugging this into the stochastic differential equation for B (t T ).8 Implementation of the Heath-Jarrow-Morton model Choose (t T ) 0 t T T (t) 0 t T : . (t T ) + 1 ( (t T ))2 2 (t) = . satisfying (7. (t T ) + 1 ( (t T ))2 2 and volatility (t T ). is 1 .CHAPTER 28.4) is equivalent to (7.3). We can then set . (t T ) and (7. above the interest rate. for every maturity T : df (t T ) = (t T ) (t T ) + (t T ) (t)] dt + (t T ) dW (t) f = (t T ) (t T ) dt + (t T ) dW (t): 28.

1) and (8. The only process which matters is (t T ) 0 t T T . it suffices to have the initial market data B (0 T T and the volatilities T: . but are usually taken to be deterministic functions.2) T and then the zero-coupon bond prices are determined by the initial conditions B (0 .1) this determines the interest rate process (8.3) in terms of W rather than f under which W W .4) (t T ). the volatility in B (t T ) must T) 0 (t T ) 0 t T In conclusion.282 These may be stochastic processes.3) f dB(t T ) = r(t)B(t T ) dt . 1 2 Zt 0 2 (u) du T) 0 T T be determined by the market. it is apparent that neither (t) nor (t T ) will enter subsequent computations. A Zt (u) dW (u) . combined with the stochastic differential equation T) 0 T (8.4) implies This is because B (T vanish. we have written (8. gotten from the market. Define (t T ) = (t T ) (t T ) + (t T ) (t) f W (t) = Zt 0 (u) du + W (t) Z f(A) = Z (T ) dIP 8A 2 F (T ): IP 0 Let f (0 Z (t) = exp . and the process (t T ) = obtained from From (8. P f is a Brownian motion.3) we see that (t T . Written this way. T ) is the volatility at time t of the zero coupon bond maturing at time (T T ) = 0 0 T T: (8.1): @ f (0 T ) = . recall from equation (4. @T log B (0 T ) 0 T T : Then f (t T ) for 0 t T is determined by the equation f df (t T ) = (t T ) (t T ) dt + (t T ) dW (t) r(t) = f (t t) 0 t T (8. ZT t (t u) du 0 t T T (8. (t T )B(t T ) dW (t): f Because all pricing of interest rate dependent assets will be done under the risk-neutral measure I .5) T ) = 1 and so as t approaches T (from below). to implement the HJM model. Equation (8.

1). In (8. and volatility is unaffected by the change of measure.3 It is customary in the literature to write W rather than W and IP rather than I . and we let B (t T ) 0 t P T T . The only parameter which must be estimated from the market is the bond volatility (t T ). @ f (0 T ) = . (t t) = @u (t u) du: t f f We then let W be a Brownian motion under a probability measure I . be given by (8. We can then define (t T ) = @ @T (t T ) and (8. Term-structure models We require that 283 (t T ) be differentiable in T and satisfy (8. @T log B (0 T ) 0 T T : . so that IP is the symbol used for the risk-neutral measure and no reference is ever made to the market measure.CHAPTER 28.2) and f (t T ) by (8.5).3).1) we use the initial conditions f f P Remark 28.4) will be satisfied because ZT @ (t T ) = (t T ) . where r(t) is given by (8.

284 .

t is the row vector t1 t2 : : : tn ]. then its distribution is determined by its mean function m(t) = IEX (t) and its covariance function (s t) = IE (X (s) . the joint density of X (t 1) : : : X (tn) is IP fX (t1) 2 dx1 : : : X (tn) 2 dxng n o 1p exp .1 (Gaussian Process) A Gaussian process X (t). where is the covariance matrix x is the row vector ::: 5 (tn t1) (tn t2 ) : : : (tn tn ) x1 x2 : : : xn]. m(t)) 1 (x . 1 (x .1 If X is a Gaussian process. the set of random variables X (t1) X (t2) : : : X (tn) is jointly normally distributed. k=1 uk X (tk ) = exp u m(t)T + 1 u 2 n uT o 285 . m(t))]: Indeed. Remark 29. and m(t) = m(t1) m(t2) : : : m(tn)]. IE exp 2 6 (t1 t1) (t1 t2) :: :: :: = 6 (t2: :t1) (t2: :t2 ) : : : 6 : 4 : ) (t1 tn ) (t2 tn ) 7 7 7 3 The moment generating function is (X n where u = u1 u2 : : : un]. is a stochastic process with the property that for every set of times 0 t1 t2 : : : tn . m(t))T dx1 : : : dxn = 2 (2 )n=2 det .Chapter 29 Gaussian processes Definition 29. t 0. m(s)) (X (t) .

286 29.) We have Therefore.1) Rs This shows that X (s) is normal with mean 0 and variance 0 2(v ) dv . a Brownian) Let dom function. (s t) = IE W (s)W (t)] = IE W (s) (W (t) .t.69 (Integral w. and shall cheat a bit by considering only two times. one must show that X (t 1) : : : X (tn) has either a density or a moment generating function of the appropriate form. W (s)) + W 2 (s) = IEW (s):IE (W (t) .g. which we usually call s and t. Zs 0 t 2(u) du: dX = dW: deuX (s) = ueuX (s) (s) dW (s) + 1 u2euX (s) 2 (s) ds 2 Zs Zs uX (s) = euX (0) + u euX (v) (v ) dW (v ) + 1 u2 euX (v) 2 (v ) dv e 2 IEeuX (s) = 1 + 1 u2 2 | 0 Zs 0 Martingale {z } 0 d IEeuX (s) = 1 u2 2(s)IEeuX (s) 2 ds Zs uX (s) = euX (0) exp 1 u2 2(v ) dv IEe 2 = exp 2(v ) dv 1 u2 2 0 2(v )IEeuX (v) dv Zs 0 (1. We shall use the m. We will want to show that IE exp fu1 X (s) + u2X (t)g = exp u1 m1 + u2 m2 + Theorem 1.f. : .. Indeed. if 0 s t.1 An example: Brownian Motion Brownian motion W is a Gaussian process with m(t) = 0 and then (s t) = s ^ t. Then Z ( 1 u1 u2 ] 2 " 11 21 12 22 # " #) u1 u2 : W (t) be a Brownian motion and (t) a nonran(u) dW (u) ^ X (t) = t is a Gaussian process with m(t) = 0 and 0 (s t) = Proof: (Sketch. W (s)) + IEW 2 (s) = IEW 2(s) = s ^ t: h i To prove that a process is Gaussian.r.

deuX (t) = ueuX (t) (t) dW (t) + 1 u2euX (t) 2 (t) dt: 2 Integrate from s to t to get Now let 0 euX (t) = euX (s) + u Take IE Zt IE Zt s : : : jF (s)] conditional expectations and use the martingale property (v )euX (v) dW (v ) F (s) = IE =0 Zt uX (v) dW (v ) + 1 u2 2 (v )euX (v) dv: (v )e 2 s s Zt 0 (v )euX (v) dW (v ) F (s) .2) (u1 +u2 )X (s) 1 (v ) dv = e exp 2 u2 2 s h i IE eu1 X (s)+u2 X (t) = IE IE eu1X (s)+u2 X (t) F (s) Z = IE e(u1 +u2 )X (s) : exp This shows that (X (s) X (t)) is jointly normal with IEX (s) = IEX (t) = 0.1) 2 (v ) dv + 1 u2 2 = exp 1 (u1 + u2 )2 2 2 2 s (v ) dv 0 Zs Zt 2 + 2u u ) 2(v ) dv + 1 u2 2 1 (u = exp 2 1 1 2 2 2 0 (v ) dv "R s 2 0R s 2# " #) ( 1 = exp 2 u1 u2 ] R0s 2 R0t 2 u1 : u2 0 0 n o Z IEX 2(s) = Zs 0 2 (v ) dv IE X (s)X (t)] = Zs 0 IEX 2(t) = Zt 0 2 (v ) dv 2(v ) dv: . for (X (s) X (t)). t 2 1 u2 2 2 s (v ) dv Zs Zt (1. Zs 0 (v )euX (v) dW (v ) to get The solution to this ordinary differential equation with initial time s is d uX (t) F (s) = 1 u2 2(t)IE euX (t) F (s) 2 dt IE e IE euX (t) F (s) = euX (s) exp 1 u2 2 IE euX (t) F (s) = euX (s) + 1 u2 2 Zt s 2 (v )IE euX (v) F (s) dv t s: t s: (1.f.g. Just as before.CHAPTER 29. where 0 s t: IE eu1 X (s)+u2X (t) F (s) = eu1 X (s)IE eu2X (t) F (s) t 2 (1.2) Zt s 2(v ) dv We now compute the m. Gaussian processes 287 s < t be given.

IE X (s)X (t)] = IE X (s)(X (t) . but note again that the Markov property follows immediately from (1. Indeed. so For fixed s Zt 0 (u) dW (u) 0. X (s) 7 6 7 4 | {z }5 = 0: 0 Therefore. Zs 2(s) = IE X (s)X (t)] = IEX IE 2(v) dv: 0 However. X (s))] = IE IE X (s)(X (t) . IE X (s)(X (t) . The equation (1. We proved this before. so its distribution is not 0 determined from its mean and covariance functions. The computation of means and variances does not require the use of moment generating functions. is to prove the normality.2 The hard part of the above argument. X is not necessarily a Gaussian process. X (s)) F (s) 2 3 6 7 = IE 6X (s) IE X (t) F (s) . the Itˆ isometry says o IE 2(v ) dv and the same argument used above shows that for 0 s t. X (t) = is a martingale and X (0) = 0.288 Remark 29. when is stochastic. the distribution of X (t) depends only R on X (s).2). Remark 29. For 0 o s t. in fact.2) says that conditioned on F (s). X (s)) + X 2(s)] Zs 2 (v ) dv: = IEX 2(s) = 0 If were a stochastic proess. m(t) = IEX (t) = 0 8t 0: IEX 2(s) = Zs 0 2(v ) dv by the Itˆ isometry.3 When is nonrandom. IEX 2(s) = Zs X (t) = Zt 0 (u) dW (u) is also Markov. X (t) is normal with mean X (s) and variance st 2(v ) dv . . and the reason we use moment generating functions.

3) Proof: (Partial) Computation of Y (s t): Let 0 s t be given. It is shown in a homework problem that (Y (s) Y (t)) is a jointly normal pair of random variables. Gaussian processes y= v= 289 z z z z s y t (a) y= s v (b) z s v y t (c) Figure 29. and let (t) and h(t) be nonrandom functions.1: Range of values of y z v for the integrals in the proof of Theorem 1. Theorem 1.3) holds. X (t) = Zt 0 z (u) dW (u) Y (t) = Zt 0 h(u)X (u) du: Then Y is a Gaussian process with mean function m Y (t) = 0 and covariance function Y (s t) = Zs 0 ^ t 2 (v ) Zs v h(y ) dy Zt v h(y) dy dv: (1. Zt 0 h(u) IEX (u) du = 0 .CHAPTER 29. Here we observe that mY (t) = IEY (t) = and we verify that (1.70.70 Let Define W (t) be a Brownian motion.

4 Unlike the process X (t) = R t (u) dW (u). 29. 29.1(a)) Zz 0 2(v ) dv = Z sZ 0 Z 0 + Zs 0 z h(y ) Zs y h(z ) dz Zy 0 2 (v ) dv = Z sZ 0 Z v + sZ y h(z ) 2 (v ) Zt z 2(v ) dv h(y ) dy dv dz 0 0 s h(z ) 2(v ) h(y ) 2(v ) Zs y Zt z h(z ) dz dv dy h(z) dz dy dv = Zs 0 + sZ s h(y ) dy dz dv 0 v Z0s 2 (v ) = 0 Zs 2 (v ) = 0 = Zs + Zs 0 2 (v ) Z sZ t v Z z h(y ) 2(v ) sZ s Zs y (See Fig.290 We have Y (s t) = IE Y (s)Y (t)] Zs Zt = IE h(y )X (y ) dy: h(z )X (z) dz = IE = = = = Z s0 Z t 0 0 Z sZ t 0 h(y)h(z )X (y)X (z ) dy dz Z0s Z0t Z0s Z0t 0 Z z h(y )h(z)IE X (y )X (z )] dy dz h(y )h(z ) Zy 0 Z ^ z 2 (v ) dv dy dz Zs 0 + sZ s 0 y h(y )h(z ) z h(z ) Zt z h(y )h(z) h(y ) dy 0 Zy 0 2 (v ) dv dy dz dz dy dz dy (See Fig. 29.1(c)) h(z ) dz dv h(y ) dy dv t Remark 29. the process Y (t) = R t X (u) du is 0 0 .1(b)) h(y )h(z) dy dz dv h(y )h(z) dz dy dv h(y )h(z) dy dz dv 2 (v ) 2 (v ) Z v y sZ t Zv s v Zv v s h(y ) dy h(y ) dy Zt Zv v (See Fig.

The conditional expectation IE not equal to Y (s). . For 0 291 IE Y (t)jF (s)] = Zs 0 s < t. nor is it a function of Y (s) alone.CHAPTER 29. h(u)X (u) du + IE Zt s = Y (s) + = Y (s) + Zt Zt s s h(u)X (u) du F (s) h(u)IE X (u) F (s)] du h(u)X (s) du = Y (s) + X (s) Zt s h(u) du Y (t)jF (s)] is where we have used the fact that X is a martingale. Gaussian processes neither Markov nor a martingale.

292 .

(t)r(t)) dt + (t) dW (t) (t) are nonrandom functions of t. 293 . K (t) = We can solve the stochastic differential equation. K (t) r(0) + Zt 0 (u) du + Zt 0 From Theorem 1. Set Zt 0 (u) du: Then d eK (t)r(t) = eK(t) (t)r(t) dt + dr(t) = eK (t) ( (t) dt + (t) dW (t)) : Integrating.Chapter 30 Hull and White model Consider where (t). we get eK (t)r(t) = r(0) + so Zt 0 eK(u) eK(u) (u) du + Zt 0 eK (u) (u) dW (u) eK(u) (u) dW (u) : r(t) = e .69 in Chapter 29. K (s) K (t) Zs 0 t 2K (u) 2 e (u) du: (0. Zt 0 ^ eK(u) (u) du (0. K (t) r(0) + .1) r (s t) = e . we see that r(t) is a Gaussian process with mean function mr (t) = e and covariance function .2) The process r(t) is also Markov. (t) and dr(t) = ( (t) .

70 in Chapter 29. K (t) dt e .r(0)C (0 T ) . r(t) dt 0 ( !) ZT ZT 1 (. K (y) dy dv C (0 T ) = A(0 T ) = ZT 0 0 Z TZt 0 e .1)2 var = exp (. K (y) dy !2 dv: The price at time 0 of a zero-coupon bond paying $1 at time T is ( ZT ) B (0 T ) = IE exp . 1 2 ZT 0 e2K (v) 2(v ) ZT v e . K (t) X (t) dt: ZT 0 r(t) = e r(t) dt = . K (t) Zt 0 eK (u) (u) du dt (0.294 RT We want to study 0 r(t) dt. 0 IE and its variance is ZT 0 r(t) dt = ZT 0 r(t) dt is normal. ZT 0 r(0) + K (t) Zt 0 eK(u) (u) du + e .3) var ZT 0 r(t) dt = IEY 2(T ) = ! ZT 0 e2K(v) 2(v ) ZT v e . 1 +2 = expf. e . we define X (t) = Then Zt 0 eK (u) (u) dW (u) Y (T ) = K (t) ZT 0 e .r(0) e K (t) dt . Its mean is r(0) + e .1)IE r(t) dt + 2 r(t) dt 0 0 ZT ZTZ t = exp . e K (t)+K (u) (u) du dt 0 0 0 !2 ZT ZT . A(0 T )g where 0 e2K (v) 2 (v ) v e . . r(0) + Zt 0 K (t) X (t) eK (u) (u) du dt + Y (T ): RT According to Theorem 1. To do this. K (y ) dy !2 dv: . K (t)+K (u) (u) du dt .

r(u) du F (t) : t 0 T ] of the zero-coupon bond: " Because r is a Markov process. Z T2 Z 4eK(v) (v) T e A(0 T ) = 0 v ZT . = dv: t t C (0 T ) = e K(y) dy 0 B(0 T ) = exp f. B (t T ) = exp f. In fact.1: Range of values of u u t for the integral. K (y) dy .r(t)C (t T ) . K (t)+K (u) (u) du dt (u) dt du K (y) dy u T eK (v) e .1 Fiddling with the formulas Note that (see Fig 30. this should be random only through a dependence on r(t). Consider the price at time t 2 ( ZT ) # B(t T ) = IE exp . K (t)+K (u) (v ) ZT v e ! .r(0)C (0 T ) . A(t T )g . A(0 T )g : .1) ZTZ t 0 = (y = t v = u) = Therefore. 1 e2K (v) 2(v ) 2 ! ZT v !23 e K(y) dy 5 dv . Z Z 0 TZ T 0 0 e . Hull and White model 295 T u Figure 30. 30.CHAPTER 30.

296 where

C (t T ) = eK(t)

Z T2 Z 4eK(v) (v) T e A(t T ) = t v ZT
t

;

K (y) dy

!

; 1 e2K(v) 2(v) 2

ZT
v

!23 e K(y) dy 5 dv
;

e

;

K (y) dy:

The reason for these changes is the following. We are now taking the initial time to be t rather than RT R zero, so it is plausible that 0 : : : dv should be replaced by tT : : : dv: Recall that

K (v ) =
and this should be replaced by

Zv
0

(u) du

K (v ) ; K (t) =

Zv
t

(u) du:

Similarly, K (y ) should be replaced by K (y ) ; K (t). Making these replacements in A(0 see that the K (t) terms cancel. In C (0 T ), however, the K (t) term does not cancel.

T ), we

30.2 Dynamics of the bond price
Let Ct(t

T ) and At (t T ) denote the partial derivatives with respect to t. From the formula

B (t T ) = exp f;r(t)C (t T ) ; A(t T )g
we have

dB (t T ) = B (t T ) ;C (t T ) dr(t) ; 1 C 2(t T ) dr(t) dr(t) ; r(t)Ct(t T ) dt ; At (t T ) dt 2
= B (t T ) ; C (t T ) ( (t) ; (t)r(t)) dt

h

i

; C (t T ) (t) dW (t) ; 1 C 2(t T ) 2(t) dt 2 ; r(t)Ct(t T ) dt ; At(t T ) dt :
Because we have used the risk-neutral pricing formula

) # ( ZT B(t T ) = IE exp ; r(u) du F (t) t "
dB(t T ) = r(t)B(t T ) dt + (: : : ) dW (t):

to obtain the bond price, its differential must be of the form

CHAPTER 30. Hull and White model
Therefore, we must have

297

;C (t T ) ( (t) ; (t)r(t)) ; 1 C 2(t T ) 2(t) ; r(t)Ct(t T ) ; At(t T ) = r(t): 2
We leave the verification of this equation to the homework. After this verification, we have the formula

dB (t T ) = r(t)B (t T ) dt ; (t)C (t T )B(t T ) dW (t): In particular, the volatility of the bond price is (t)C (t T ).
30.3 Calibration of the Hull & White model
Recall:

C (t T ) = eK(t) e K(y) dy t B (t T ) = exp f;r(t)C (t T ) ; A(t T )g :
;

ZT2 Z 4eK(v) (v) T e A(t T ) = t v ZT
0

dr(t) = ( (t) ; (t)r(t)) dt + (t) dB(t) Zt K (t) = (u) du
;

K (y)

dy ; 1 e2K (v) 2 (v ) 2

!

ZT
v

!2 3 e K (y) dy 5 dv
;

Suppose we obtain B (0 determine the functions

T ) for all T 2 0 T ] from market data (with some interpolation). Can we (t), (t), and (t) for all t 2 0 T ]? Not quite. Here is what we can do.
;

We take the following input data for the calibration: 1. 2. 3. 4. 5.

B (0 T ) 0 T T r(0);
(0); (t) 0 t T

(usually assumed to be constant);

(0)C (0 T ) 0 T

T

, i.e., the volatility at time zero of bonds of all maturities.

Step 1. From 4 and 5 we solve for

C (0 T ) =

ZT
0

e

;

K (y)

dy:

298 We can then compute

We now have

@ C (0 T ) = e K(T ) @T @ =) K (T ) = ; log @T C (0 T ) @ K (T ) = @ Z T (u) du = (T ): @T @T 0 (T ) for all T 2 0 T ].
;

Step 2. From the formula

B(0 T ) = expf;r(0)C (0 T ) ; A(0 T )g we can solve for A(0 T ) for all T 2 0 T ]. Recall that

Z T2 ZT A(0 T ) = 4eK (v) (v ) e
0

v

;

K (y) dy

!

; 1 e2K(v) 2(v) 2
T
as follows:

ZT
v

!23 e K (y) dy 5 dv:
;

We can use this formula to determine

ZT @ A(0 T ) = K (v) (v )e K (T ) ; e2K (v) 2 (v )e K (T ) e e K (y) dy dv @T 0 v !# Z " ZT K (T ) @ A(0 T ) = T eK (v) (v ) ; e2K (v) 2 (v ) K (y) dy e @T e dv 0 v @ eK (T ) @ A(0 T ) = eK (T ) (T ) ; Z T e2K(v) 2(v ) e K (T ) dv @T @T 0 @ eK (T ) @ A(0 T ) = e2K (T ) (T ) ; Z T e2K(v) 2 (v ) dv eK(T ) @T @T 0 @ eK(T ) @ eK (T ) @ A(0 T ) = (T )e2K (T ) + 2 (T ) (T )e2K(T ) ; e2K (T ) 2(T ) 0 T T : @T @T @T
; ; ; ; ; 0 0

(T ) 0 T

ZT"

!#

This gives us an ordinary differential equation for , i.e.,

(t)e2K (t) + 2 (t) (t)e2K (t) ; e2K (t) 2(t) = known function of t:

From assumption 4 and step 1, we know all the coefficients in this equation. From assumption 3, we have the initial condition (0). We can solve the equation numerically to determine the function (t) 0 t T . Remark 30.1 The derivation of the ordinary differential equation for (t) requires three differentiations. Differentiation is an unstable procedure, i.e., functions which are close can have very different derivatives. Consider, for example,

f (x) = 0 8x 2 IR g (x) = sin(1000x) 8x 2 IR: 100

CHAPTER 30. Hull and White model
Then

299

1 jf (x) ; g(x)j 100 8x 2 IR
but because

g (x) = 10 cos(1000x)
0

we have for many values of x.

jf (x) ; g (x)j = 10
0 0

Assumption 5 for the calibration was that we know the volatility at time zero of bonds of all maturities. These volatilities can be implied by the prices of options on bonds. We consider now how the model prices options.

30.4 Option on a bond
Consider a European call option on a zero-coupon bond with strike price K and expiration time T 1 . The bond matures at time T2 > T1. The price of the option at time 0 is R T1 r(u) du ; +

R T1 r(u) du 0 (expf;r(T1)C (T1 T2) ; A(T1 T2)g ; K )+ : Z Z + = e x expf;yC (T1 T2) ; A(T1 T2)g ; K f (x y) dx dy RT where f (x y ) is the joint density of 0 1 r(u) du r(T1) . R We observed at the beginning of this Chapter (equation (0.3)) that 0T1 r(u) du is normal with "Z T1 # Z T1
= IEe
; 1 1 ; ;1 ;1

IE e

0

(B (T1 T2) ; K )

1 = IE
4

0

r(u) du =
=

Z T1
0

0

IEr(u) du r(0)e
;

2 = var 1
4

"Z T1
0

r(u) du =

# Z0T1

K (v) + e K (v)
;

Zv
0

e2K (v) 2(v )

Z T1
v

eK(u) (u) du dv

e

;

K (y) dy

!2

dv:

We also observed (equation (0.1)) that r(T1) is normal with
4 ; ;

Z T1 = IEr(T1) = r(0)e K (T1) + e K (T1 ) eK (u) (u) du 2
;

2 = var (r(T )) = e 2K (T1 ) 1 2
4

Z T1
0

0 2K (u) 2 (u) du: e

300 In fact,

R T1 r(u) du r(T ) 1 0

is jointly normal, and the covariance is

1 2 = IE

"Z T1
0

= =

Z T1 Z T1
0 0

(r(u) ; IEr(u)) du: (r(T1) ; IEr(T1))

#

IE (r(u) ; IEr(u)) (r(T1) ; IEr(T1))] du
r (u T1) du

where r (u

T1) is defined in Equation 0.2.

The option on the bond has price at time zero of

Z

1

Z

1

;1

;1

e

;

x

expf;yC (T1 T2) ; A(T1 T2)g ; K 2
1

+

1 p 2 1;

2 2 exp ; 2(1 ; 2) x2 + 2 xy + y 2 2 1 2 1 2

(

1

"

#)

dx dy:

(4.1)

The price of the option at time t 2

0 T1] is

IE e

;

R T1 r(u) du
t

(B (T1 T2) ; K )+ F (t)
;

= IE e

R T1 r(u) du
t

(expf;r(T1)C (T1 T2) ; A(T1 T2)g ; K )+ F (t)

(4.2)

Because of the Markov property, this is random only through a dependence on r(t). To compute R this option price, we need the joint distribution of tT1 r(u) du r(T1) conditioned on r(t). This

CHAPTER 30. Hull and White model
pair of random variables has a jointly normal conditional distribution, and

301

1 (t) = IE

"Z T1
t

2 Z 3 !2 T1 2 (t) = IE 4 r(u) du ; 1(t) F (t)5 1 t !2 Z T1 Z T1
=
t

=

Z T1
t

r(u) du F (t)
;

#

r(t)e

K (v)+K (t) + e K (v)
;

Zv
t

eK (u) (u) du dv

e2K (v) 2(v )

v

e

;

K (y) dy

dv

2 (t) = IE

Z T1 = r(t)e K (T1 )+K (t) + e K (T1 ) eK(u) (u) du
; ;

r(T1) r(t)

2 (t) = IE 2

(r(T1) ;
2K (T1 )

=e

;

Z T1
t

2

(t))2

F (t) !

t

(t) 1 (t) 2 (t) = IE =
t

" Z T1
t

e2K (u) 2 (u) du

Z T1

e

;

Z K (u) K (T1 ) u e2K (v) 2 (v ) dv du: t
;

r(u) du ; 1 (t) (r(T1) ; 2(t)) F (t)

#

The variances and covariances are not random. The means are random through a dependence on r(t). Advantages of the Hull & White model: 1. Leads to closed-form pricing formulas. 2. Allows calibration to fit initial yield curve exactly. Short-comings of the Hull & White model: 1. One-factor, so only allows parallel shifts of the yield curve, i.e.,

B (t T ) = exp f;r(t)C (t T ) ; A(t T )g
so bond prices of all maturities are perfectly correlated. 2. Interest rate is normally distributed, and hence can take negative values. Consequently, the bond price " ( ) #

B (t T ) = IE exp ;

ZT
t

r(u) du F (t)

can exceed 1.

302

Since. For j = 1 : : : d. r(t) is a Gaussian process. 1 2 t Xj (0) + 1 2 . 1 Xj (t) dt + 1 dWj (t): 2 2 Xj is called the Orstein-Uhlenbeck process. Let > 0 and > 0 be 2 2 2 X1 (0) + X2 (0) + : : : + Xd (0) 0 and let Xj be the solution to the stochastic differential equation dXj (t) = . IP fr(t) > 0g = 1. there is a positive probability that r(t) < 0. The Cox-Ingersoll-Ross model is the simplest one which avoids negative interest rates. r(t) is normally distributed. but (see Fig. for each t. 31. We begin with a d-dimensional Brownian motion (W 1 W2 constants. It always has a drift toward the origin. Zt 0 e 2 u dWj (u) : 1 This solution is a Gaussian process with mean function mj (t) = e and covariance function 1 2 t Xj (0) ^ (s t) = 1 2 e 4 4 .1) IP There are infinitely many values of t > 0 for which r(t) = 0 303 = 1 .Chapter 31 Cox-Ingersoll-Ross model In the Hull & White model. we have r(t) = X1 (t) and for each t. The solution to this stochastic differential equation is Xj (t) = e . let Xj (0) 2 IR be given so that : : : Wd ) . 1 (s+t) Z s t u 2 e 0 du: Define 2 2 2 r(t) = X1 (t) + X2 (t) + : : : + Xd (t): 2 If d = 1.

r(t) dt + 2 d q XX = d4 . Then 1 2 d fxi = 2xi fxi xj = 2 0 d X i=1 ( if i = j if i 6= j: Itˆ ’s formula implies o dr(t) = = d X i=1 d X i=1 fxi dXi + 1 2 fxi xi dXi dXi d X1 i=1 4 2 2Xi . 31. If d 2. r(t) dt + r(t) p i(t) dWi(t): i=1 r(t) Define Xi dWi + d4 dt 2 d X Z t Xi(u) p dWi(u): W (t) = r(u) i=1 0 . 1 Xi dt + 1 dWi(t) + 2 2 d X i !=1 dWi dWi = .1) IP fThere is at least one value of t > 0 for which r(t) = 0g = 0: Let f (x1 x2 : : : xd) = x2 + x2 + : : : + x2 .304 r(t) = X 1 (t) 2 t x2 ( X (t). (see Fig.1: r(t) can be zero. X (t) ) 2 1 x 1 Figure 31.

1. We have ! d 2 . one can derive formulas under the assumption that integer. Xi (t) is normal with mean zero and variance 2 (t t) = 4 (1 . Let r(0) . here is the distribution of r(t) for fixed t > 0.e. then 2 IP fThere is at least one value of t > 0 for which r(t) = 0g = 0: With the CIR process.e. d= 4 2 is a positive 0 be given. IP fThere are infinitely many values of t > 0 for which r(t) = 0g = 1: This is not a good parameter choice. e t ): . r(t)) dt + The Cox-Ingersoll-Ross (CIR) process is given by q r(t) dW (t) We define If d happens to be an integer. 305 d X Xi dW = pr dWi dW dW = d X Xi2 i=1 i=1 r dt = dt so W is a Brownian motion. For example. If d 2 (i. then we have the representation d = 4 2 > 0: d X i=1 r(t) = Xi2(t) <1 2 2). and they are still correct even when d is not an integer. .. 1 2 ).CHAPTER 31. r(t) dt + qr(t) dW (t): dr(t) = 4 dr(t) = ( . then but we do not require d to be an integer. Take X1(0) = 0 X2(0) = 0 : : : Xd 1(0) = 0 Xd(0) = r(0): For i = 1 q 2 : : : d . If d < 2 (i. Cox-Ingersoll-Ross model Then W is a martingale..

2 2 2 4 r 2 2 .7.2 Normal squared and independent of the other 2 Xd (t) | {z } (0. 31. and so the limiting distribution of r(t) is 4 times a chi-square with d = 4 2 degrees of freedom.1 Equilibrium distribution of r (t) As t!1. 2 2 e .2 Kolmogorov forward equation Consider a Markov process governed by the stochastic differential equation dX (t) = b(X (t)) dt + (X (t)) dW (t): . md (t)!0. 1 22 r 2 . 1 q 2 t r(0) r(t) = Chi-square with d . Then md(t) = e d 1 X . 1 = freedom | (t t) i=1 {z X p i((tt)t) 4 !2 } degrees of + term . 2 2 22 y 2 . 31. 2 2 e . 2 2 r: We computed the mean and variance of r(t) in Section 15. 2 2 r = 22 . we have (t t) = 4 .306 Xd (t) is normal with mean and variance (t t). We have d X Xi(t) !2 p r(t) = (t t) : (t t) i=1 2 2 As t!1. y=2 : y. The limiting density for r(t) is 2 2 p(r) = 4 2 : 1 2 = 2.2 2 e . . The chi-square density with 4 2 degrees of freedom is f (y ) = We make the change of variable r = 4 1 2 = 2.1) 2 Thus r(t) has a non-central chi-square distribution.

Since 0 and x will not change during the following. Z 1 0 h(y )p(t y ) dy t and y .2). We have IEh(X (t)) = for any function h. we assume that 0 at time 0. X (t) = r(t). we omit them and write p(t y ) rather than p(0 t x y ). Cox-Ingersoll-Ross model 307 h 0 -y Figure 31. Let h(y ) be a smooth function of y 0 which vanishes near y = 0 and for all large values of y (see Fig.2: The function h(y ) Because we are going to apply the following analysis to the case X (t) 0 for all t.CHAPTER 31. 31. We derive it below. Itˆ ’s formula implies o The Kolmogorov forward equation (KFE) is a partial differential equation in the “forward” variables dh(X (t)) = h (X (t))b(X (t)) + 1 h (X (t)) 2(X (t)) dt + h (X (t)) (X (t)) dW (t) 2 0 00 0 h i so h(X (t)) = h(X (0)) + Z th 0 0 Zt 0 h (X (s))b(X (s)) + 1 h (X (s)) 2 (X (s)) ds + 2 0 00 i h (X (s)) (X (s)) dW (s) IEh(X (t)) = h(X (0)) + IE Z th 0 h (X (s))b(X (s)) dt + 1 h (X (s)) 2 (X (s)) ds 2 0 00 i . Then X (t) is random with density p(0 t x y ) (in the y We start at X (0) = x variable).

Z 0 1 h(y)pt(t y) dy = . It implies that 2 (y )p(t y) = 0: (KFE) If there were a place where (KFE) did not hold. Z 0 1 h(y)p(t y) dy = h(X (0)) + Z tZ 1 1 Zt 1 2 0 0 Z0 0 h (y )b(y )p(s y) dy ds + 0 h (y ) 2 (y)p(s y) dy ds: 00 Differentiate with respect to t to get Z 1 0 h(y)pt(t y) dy = Z 1 0 h (y )b(y )p(t y ) dy + 1 2 0 Z 0 1 h (y ) 2 (y)p(t y ) dy: 00 Integration by parts yields Z 0 1 h (y)b(y)p(t y ) dy = h(y)b(y )p(t y ) 0 y= 1 Z 0 | 1 h (y ) 00 2(y )p(t y) dy = h (y) 0 | 2 (y )p(t =0 =0 {z y=0 } y= . Z 0 @ h (y ) @y 0 2(y )p(t y ) dy 2(y )p(t @ = . 1 Z 0 1 @ h(y ) @y (b(y )p(t y)) dy 1 {z y) y=0 } . but take h to be zero elsewhere.308 or equivalently. 1 @y2 ( 2p) dy 6= 0: 2 " # .2. 1 @y 2 2 " 2 (y )p(t y) # dy = 0: This last equation holds for every function h of the form in Figure 31. and we would obtain h(y ) > 0 at that and nearby Z 1 0 @ @2 h pt + @y (bp) . Z 0 1 @ @2 h(y) pt(t y ) + @y (b(y )p(t y )) . 1 @y2 2 @ @2 pt(t y) + @y ((b(y )p(t y )) .h(y ) @y | 2(y )p(t =0 {z y) y= 1 y=0 } + Z 0 1 @2 h(y) @y 2 y ) dy: Therefore. then we could take points. Z 0 1 Z @ @2 h(y ) @y (b(y )p(t y )) dy + 1 h(y ) @y2 2 0 1 2(y )p(t y ) dy or equivalently.

we obtain the equilibrium Kolmogorov forward equation 0 = t! pt (t y ): lim 1 @ (b(y)p(y )) . r p(r) 0 .2 2 e . it will be 309 p(y ) = t! p(t y ): lim 1 In order for this limit to exist. : p(rr) . Cox-Ingersoll-Ross model If the process X (t) has an equilibrium density. + 2 ( . )p(r) + 2r . 2 .CHAPTER 31. 22r2 00 2 p (r) = 2 . 1 @ 2 2 @y 2 @y 2 (y )p(y ) = 0: When an equilibrium density exists. r) . r)p(r)) . 1 2 . 2 2 p(r) 2 2 1 2 = 2 r . 1 2 . r)2 p(r) = 2r r 2 2 2r 0 We want to verify the equilibrium Kolmogorov forward equation for the CIR process: @ (( . r p(r) + 2r (. r p (r) 2 2 2 2 2 . 1 2 . it is the unique solution to this equation satisfying p(y ) 0 8y 0 Z 1 0 p(y ) dy = 1: 31.3 Cox-Ingersoll-Ross equilibrium density We computed this to be p(r) = Cr where 2 . 1 22 : p (r) = . 1 @ 2 ( 2 rp(r)) = 0: 2 @r2 @r (EKFE) . 1 ( . 1 2 . we must have Letting t!1 in (KFE). 2 2 r C = 22 We compute 2 2 .

The Markov property implies that B (t T ) is random only through a dependence on r(t). 2r ( . 1 2 . r(u) du B (t T ) = IE exp . 1 2 . 1 2 2 ( . 1 2rp (r) 2 2 ( . 1 2 . B(t T ) = B(r(t) t T ): . there is a function B (r t T ) of the three dummy variables r t T such that the process B (t T ) is the function B (r t T ) evaluated at r(t) t T .4 Bond prices in the CIR model The interest rate process r(t) is given by dr(t) = ( . 2 r ( 2 2 1 . r) 2 ( . r)p(r)) = . r . r) . r)p (r) . r) = p(r) . r(u) du F (t) : t " Because ( ZT ) # exp . r)2 2 2 2 = p(r) ( .. p(r) + ( . i. + ( .310 Now @ (( . r)p (r) @r @ 2 ( 2 rp(r)) = @ ( 2 p(r) + 2rp (r)) @r2 @r = 2 2 p (r) + 2rp (r): 0 0 0 00 The LHS of (EKFE) becomes . The bond price process is q r(t) dW (t) ( ZT ) # B(t T ) = IE exp . 1 2 . 1 2 . r) =0 1 + r( 2 2 2r . r) + . 1 2 . r(u) du F (t) Zt 0 0 " the tower property implies that this is a martingale. Thus. r) 2 2 2r . 2 ) 2 r 2 0 0 00 1 + r ( . 1 2 . r)2 as expected. 31. r(t)) dt + where r(0) is given. 2p (r) . p(r) + ( .e. 2 2 .

r)Br (r t T ) + 1 2rBrr (r t T ) = 0 2 The terminal condition is 0 t < T r 0: (4. Then we have Bt = (. We comZt Z0t 0 d exp . this equation has a closed form solution. where C (T T ) = 0 A(T T ) = 0. B (A t + C ) = rB (.1 .rCt .1 .CHAPTER 31.rB + (. ( . we look for a solution of the form .r(t)B(r(t) t T ) dt + Br (r(t) t T ) dr(t) + t T ) dr(t) dr(t) + Bt (r(t) t T ) dt : 1 Brr (r(t) 2 The expression in : : : ] equals = . r) dt + Br 1 + 2 Brr 2 r dt + Bt dt: pr dW Setting the dt term to zero. Surprisingly. Using the Hull & White model as a guide. 0 r(u) du B(r(t) t T ) is a martingale.CB Brr = C 2B and the partial differential equation becomes 0 = . r(u) du B (r(t) t T ) r(u) du .1) B(r T T ) = 1 r 0: B (r t T ) = e rC (t T ) A(t T ) . r)CB + 1 2 rC 2B 2 1 2 C 2 ) . we obtain the partial differential equation . Ct + C + 2 We first solve the ordinary differential equation . = exp . Cox-Ingersoll-Ross model Because exp pute 311 n Rt o .rB dt + Br ( . At )B . At)B Br = . rB(r t T ) + Bt (r t T ) + ( . its differential has no dt term.rCt . Ct(t T ) + C (t T ) + 1 2C 2(t T ) = 0 C (T T ) = 0 2 and then set A (t T ) = ZT t C (u T ) du .

A( )g 0 r 0 where q C ( ) = cosh( sinh( 1 )sinh( ) )+ 2 2 1 2 log 4 e2 A( ) = . Moreover. r(u) du : T ) is strictly decreasing in T . we have ( ZT ) # IE exp .rC (t T ) . the function B (r t T ) depends on t and T only through their difference = T . where = 1 2 q 2+2 2 u u sinh u = e . so B (r(0) B (r(0) 0) = 1 . t. e 2 .312 so A(T T ) = 0 and At (t T ) = . C (t T ): It is tedious but straightforward to check that the solutions are given by C (t T ) = sinh( (T . 2 cosh( (T .rC ( ) . almost surely. Because the coefficients in dr(t) = ( . and we have B(r ) = exp f. We write B(r ) instead of B (r t T ). t)) + 1 sinh( (T . A(t T )g 0 t < T r 0 and C (t T ) and A(t T ) are given by the formulas above. t. C (t T ) and A(t T ) are functions of = T . t)) cosh( (T . Similarly. 0 Now r(u) > 0 for each u. t)) 2 2 3 1 2 log 4 e 2 (T t) 5 A(t T ) = . u u cosh u = e + e : 2 . = We have 1 2 q 2 cosh( ) + 1 sinh( ) 2 3 5 2 + 2 2: ( ZT ) B(r(0) T ) = IE exp . t)) + 1 sinh( (T . r(u) du F (t) = B (r(t) t T ) t " where B(r t T ) = exp f. t)) 2 . Thus in the CIR model. r(t)) dt + r(t) dW (t) do not depend on t.

T 2 is the maturity time of the bond. A(T )g r(0)C (0) + A(0) = 0 lim r(0)C (T ) + A(T )] = 1 T! 1 and r(0)C (T ) + A(T ) is strictly inreasing in T . and 0 t T 1 n R o T2. 31. As usual.r(0)C (T ) .5 Option on a bond The value at time t of an option on a bond in the CIR model is ( Z T1 ) # + F (t) v(t r(t)) = IE exp .rv + vt + ( .) The terminal condition is . K ) t " where T1 is the expiration time of the option. K )+ r 0: Other European derivative securities on the bond are priced using the same partial differential equation with the terminal condition appropriate for the particular security. r)vr + 1 2rvrr = 0 0 t < T1 r 0: 2 v(T1 r) = (B (r T1 T2) . so B(r(0) T ) = exp f. r(t)) dt + r(t) dW (t): Real time scale: In this time scale. the interest rate r( t) is given by a time-dependent CIR equation ^^ ^ ^r ^ ^ ^ ^ ^ ^ ^ dr(t) = (^(t) . the interest rate r(t) is given by the constant coefficient CIR equation q dr(t) = ( . exp . r(u) du (B (T1 T2) . Cox-Ingersoll-Ross model 313 lim B (r(0) T ) = IE exp . ^(t)^(t)) dt + ^ (t) r(t) dW (t): ^^ q .CHAPTER 31. 31. and this leads to the partial differential equation (where v = v (t r). 0t r(u) du v(t r(t)) is a martingale.6 Deterministic time change of CIR model Process time scale: In this time scale. T! 1 Z 0 1 r(u) du = 0: But also.

. . There is a strictly increasing time change function t Fig. . .t: 314 6 Process time . . . . . . .3). . . r('(t))' (t) dt 0 0 and this will be B (^(0) r ^ 0 T ) if we set ^ ^ r(t) = r('(t)) ' (t): ^^ 0 . . We want to set things up so t ^ ^rt ^ B (^ ^ T ) = B (r t T ) = e . . rC (t T ) A(t T ) . . . . . dt = ' (t) dt in the first integral to get 0 ( Z T^ ) ^ ^ ^ B (r(0) 0 T ) = IE exp . . . . . . . . . . ^ = '( t) which relates the two time scales (See ^rt ^ ^ ^ Let B (^ ^ T ) denote the price at real time t of a bond with maturity T when the interest rate at time ^ is r. We have ( ZT ) B(r(0) 0 T ) = IE exp . . ^ t = '(t) - ^ t: Real time Figure 31. . r(t) dt : r ^^ ^ 0 With T ^ ^ ^ ^ = '(T ). . . . . . . . . . ^ where t = '(t) ^ T = '(T ).3: Time change function. and C (t T ) and A(t T ) are as defined previously. . . . r(t) dt ( Z0T ) ^ ^ B(^(0) 0 T ) = IE exp . . . . . . . . . . . ^ We need to determine the relationship between r and r. 31. . . make the change of variable t = '(t). . . . . . A period of high interest rate volatility - . .

We calibrate by writing the equation ^r ^ B (^(0) 0 T ) = exp . Cox-Ingersoll-Ross model 315 31. since.CHAPTER 31. Take '0(0) = 1 (we justify this in the next section). the model coefficients ^ ^ ^ 0 T ) for all T 2 0 T ]. (*) 0 and ^r ^ ^ ^ Since 0 .. i. A(t T ) ^^^ ^ ^ ^^ ^ 0 0 ! where ^ '^ ^^^ C (t T ) = C ('(t) ^ (T )) ' (t) ^^^ ^ ^ A(t T ) = A('(t) '(T )) 0 ^ do not depend on ^ and T only through t are time dependent. in the real time scale. For T ^1 < T2.r(t) ^^ ^ ' (t) n o = exp .e. A('(t) '(T )) ^ ^ = exp . starting at 0 at time having limit 1 at 1. this solution ^ ^ so '(T1) < '(T2). t. For each T . Thus ' is a strictly increasing time-change-function with the right properties.r(t)C (t T ) . log B (^(0) 0 T ) = '(0) C ('(0) '(T )) + A('(0) '(T )): (0) 0 Take and so the equilibrium distribution of r(t) seems reasonable. log B (^(0) 0 T ) < 1 (*) has a unique solution for each T . log B (^(0) 0 T ) = r(0)C (0 '(T )) + A(0 '(T )): ^ The right-hand side of this equation is increasing in the '( T ) variable. If T ^ ^ ^ ^ . n o ^ ^ ^ ^r ^ r . . then ^ is '(0) = 0. ^r ^ Suppose we know r(0) and B (^(0) ^ ^ T . log B (r(0) 0 T1) < .7 Calibration ^ ^) ^ ^ ^r^ t ^ B (^(t) ^ T ) = B r(t^ '(t) '(T ) ' ) ( (t) ^ ^ C ('(t) '(T )) . These values determine ^ the functions C A. solve the ^): equation for '(T ^r ^ ^ ^ ^ . A(0 T ) ^ ^ ^ ^ ^ or equivalently.r(0)C (0 T ) . log B (r(0) 0 T2) r(0)C (0 0) + A(0 0) = 0 ^ lim r(0)C (0 T ) + A(0 T )] = 1: ^ T! 1 = 0.

2 2 log 4 cosh( T ) e 1 sinh( T ) 5 +2 = 1 2 q 2 + 2 2: . log B (^(0) 0 T ) = ' 1 r(0)C ('(T )) + A('(T )) T 0 (0) ^ 0 (cal) = 0. r(0)C (T ) + A(T ) ^ C (T ) = sinh( T ) cosh( T ) + 1 sinh( T ) 2 2 3 1 T 2 A(T ) = . ( ZT ) B(0 T ) = IE exp . log IE exp . . r(u) du 0 R T r(u) du T =0 In the real time scale associated with the calibration of CIR by time change. . log B(0 T ) = . @T log B(0 T ) Justification: T =0 = r(0): @ log B (0 T ) = IE r(T )e 0 . Consider the function Here C (T ) and A(T ) are given by ^r ^ ^ ^ ^ .316 31. @T log B(0 T ) = r(0): .4. is strictly increasing in T .8 Tracking down '0 (0) in the time change of the CIR model Result for general term structure models: @ . 31. @^ log B(^(0) 0 T ) ^ = r(0): ^ @T T =0 The calibration of CIR by time change requires that we find a strictly increasing function ' with '(0) = 0 such that ^r ^ ^ ^ where B (^(0) 0 T ). r(u) du : 0 ( ZT ) . Therefore. @T RT IEe 0 r(u) du @ . The above says that ^r ^ . determined by market data. log B (^(0) 0 T ) is as shown in Fig. starts at 1 when T ^r ^ ^ and goes to zero as T !1. we write the bond price as ^r ^ B(^(0) 0 T ) thereby indicating explicitly the initial interest rate.

.4: Bond price in CIR model 6 r(0)C (T ) + A(T ) ^ ^r ^ ... .. . .. ^ '(T ) Figure 31.. .. except it satisfies (cal’) rather than (cal). . ... . . then '(T ) = 0. . is strictly increasing in T .. ^r ^ ^ . . .. . As T !1. .5: Calibration -T ^ The function r(0)C (T ) + A(T ) is zero at T = 0. '(T )!1.... . . . .. log B (^(0) 0 T ) .. log B (^(0) 0 T ) = r(0)C (T ) + A(T ) ^ ^ ^ ^ ^ ^ ^ ^ If T = 0.... . . log B (^(0) 0 T ) Strictly increasing -T ^ Figure 31.... then '(T1) < '(T2)... ... ..... Cox-Ingersoll-Ross model 317 Goes to 1 ^r ^ 6... .4).5) (cal’) defined a time-change function ' which has all the right properties. . . and goes to 1 as T !1. 31...CHAPTER 31. . . let us first consider the related equation ^r ^ ^ ^ ^ ... .. . . .. ... If T1 < T2... .. log B (^(0) 0 T ) = r(0)C ('(T )) + A('(T )): ^ ^ Fix T and define '(T ) to be the unique T for which (see Fig...... .. We have thus .. To solve (cal)... .. .. . This is because the interest rate is positive in the CIR model (see last paragraph of Section 31. . .... ...

From (cal’) we compute ^r ^ r(0) = . @^ log B (^(0) 0 T ) ^ ^ @T T =0 = r(0)C ('(0))' (0) + A ('(0))' (0) ^ = r(0)C (0)' (0) + A (0)' (0): ^ 0 0 0 0 0 0 0 0 We show in a moment that C 0 (0) = 1. so we have r(0) = r(0)' (0): ^ ^ 0 Note that r(0) is the initial interest rate. 0(0 + 1 ) = 1: 2 0 .318 We conclude by showing that ' 0(0) = 1 so ' also satisfies (cal). observed in the market. and is striclty positive. e A (0) = . Dividing by ^ r(0). (0 + 2 ) " 2 1 2 . A0(0) = 0.2 2 0 " cosh( ) + 1 sinh( ) 2 # 2 1 cosh( ) + 1 sinh( ) 2 e =2 =2 2e 1 cosh( ) + 2 sinh( ) . we obtain ^ ' (0) = 1: 0 Computation of C 0(0): C( )= 0 1 cosh( ) + 1 sinh( ) 2 2 cosh( ) )+ 1 2 1 cosh( ) + 2 sinh( ) h i C (0) = 12 ( + 0) . sinh( ) 2 sinh( cosh( ) Computation of A0 (0): A ( ) = .2 2 1 )+ 2 cosh( ) . 2 2 12 = 0: 1 1 ( + 0)2 #2 ( + 0) . 2 2 0 =2 2 sinh( +0 = .

2W2 (t): 4 4 Then W3 is a Brownian motion with dW1(t) dW3(t) = dt and dX1 dX1 = 2Y dt dX2 dX2 = 2Y dt dX1 dX2 = 1 2 319 1 2Y dt: . To simplify notation. 2 dW2(t)) (SDE2) where W1 and W2 are independent Brownian motions.Chapter 32 A two-factor model (Duffie & Kan) Let us define: X1 (t) = Interest rate at time t X2(t) = Yield at time t on a bond maturing at time t + Let X1(0) > 0. X2 (0) differential equations 0 > 0 be given. and let X1(t) and X2(t) be given by the coupled stochastic dX1(t) = (a11X1(t) + a12X2(t) + b1) dt + 1 1X1(t) + 2X2(t) + dW1(t) (SDE1) q q dX2(t) = (a21X1(t) + a22X2(t) + b2) dt + 2 1X1(t) + 2X2(t) + ( dW1(t) + 1 . we define q Y (t) = 1X1(t) + 2X2 (t) + q W3(t) = W1(t) + 1 .

2. ) dt 2 2 2 2 1p 1 1 + 2 1 2 1 2 + 2 2 ) 2 Y dW4 2 2 2 2 1p 1b1 + 2 b2 . 2 2 dW2) = ( 1a11 + 2 a21)X1 + ( 1a12 + 2a22)X2 ] dt + ( 1 b1 + 2 b2) dt q 2 2 1 + ( 1 2 + 2 1 2 1 2 + 2 2 ) 2 Y (t) dW4(t) 1 2 1 1 + 2 2 )W1(t) + 2 1 . we recall that Y will stay strictly positive provided that: Assumption 2: and Assumption 3: Under Assumptions 1. ) dt + ( 1 1 + 2 1 2 1 2 + 2 2) 2 Y dW4: From our discussion of the CIR process. We shall choose the parameters so that: Assumption 1: For some . and 3. 2 2 W2 (t) q2 2 2 2 1 1+2 1 2 1 2+ 2 2 where W4(t) = ( p is a Brownian motion. Y (0) = 1X1 (0) + 2X2(0) + > 0 1( 2 2 + 2 1 2 2 1 1 2 2 1 2 + 2 2 ): 1b1 + 2 b2 . These can be rewritten as dX1(t) = (a11X1 (t) + a12X2 (t) + b1) dt + 1 Y (t) dW1(t) q dX2(t) = (a21X1 (t) + a22X2 (t) + b2) dt + 2 Y (t) dW3(t): q (SDE1’) (SDE2’) .1 Non-negativity of Y dY = 1 dX1 + 2 dX2 = ( 1a11X1 + 1a12 X2 + 1b1) dt + ( 2a21 X1 + 2a22 X2 + 2 b2) dt q p + Y ( 1 1 dW1 + 2 2 dW1 + 2 1 .320 32. Y (t) > 0 0 t < 1 almost surely. Then 1 a11 + 2a21 = 1 1 a12 + 2a22 = 2: dY = 1X1 + +( = Y dt + ( 2X2 + ] dt + ( 1b1 + 2b2 . and (SDE1) and (SDE2) make sense.

X1B + (a11X1 + a12X2 + b1)Bx1 + (a21X1 + a22X2 + b2)Bx2 . = exp . B p 1 2 Y Bx1 x2 + 1 2 Y Bx2 x2 2 2 2 + 1 2 Y Bx1 x1 + 2 1 1 Y Bx1 dW1 + p dt Y Bx2 dW3 The partial differential equation for B (x 1 x2 ) is . B +(a11x1 + a12x2 + b1)Bx1 +(a21x1 + a22x2 + b2)Bx2 + 1 2( 1x1 + 2x2 + )Bx1 x1 2 1 1 + 1 2( 1x1 + 2x2 + )Bx1 x2 + 2 2 ( 1x1 + 2x2 + )Bx2 x2 = 0: (PDE) 2 B (x1 x2 ) = exp f. t) is a martingale. t) X1(u) du . this is random only through a dependence on X1(t) X2(t). Zt 0 X1(u) du B(X1(t) X2(t) T .2 Zero-coupon bond prices The value at time t T of a zero-coupon bond paying $1 at time T is ( ZT ) # B(t T ) = IE exp . B dt + 1 Bx1 x1 dX1 dX1 + Bx1 x2 dX1 dX2 + 1 Bx2 x2 dX2 dX2 2 2 X1(u) du + . X1(u) du F (t) : t " The usual tower property argument shows that exp . there is a function B (x 1 x2 ) of the dummy variables x1 x2 and . Zt 0 Z0t Zt 0 X1(u) du B(X1(t) X2(t) T .x1 C1 ( ) . t. the bond price depends on t and T only through their difference = T . Thus. A( )g 0 and all x1 x2 satisfying 1x1 + 2 x2 + > 0: We seek a solution of the form valid for all (*) . t) = IE exp .X1 B dt + Bx1 dX1 + Bx2 dX2 . x2C2( ) . x1B . X1(u) du F (t) : t " Since the pair (X1 X2) of processes is Markov. = exp . We compute its stochastic differential and set the dt term equal to zero. d exp . Since the coefficients in (SDE1) and (SDE2) do not depend on time.CHAPTER 32. so that ( ZT ) # B(X1 (t) X2(t) T . A two-factor model (Duffie & Kan) 321 32.

a22C2 ( ) 2 C 2( 1 2 1 0 )+ 1 2 2C1( 2 C 2( 2 2 2 + B (x1 x2 ) A ( ) . 1 2 1 A(0) = 0 0 1 2 2C1( 2 )C2( ) .322 We must have B(x1 x2 0) = 1 8x1 x2 satisfying (*) because = 0 corresponds to t = T . 1 2 2 C 2( 2 1 2 ) (1) (2) (3) C1(0) = 0 2 C2( ) = a12C1( ) + a22 C2( ) .x1 C1( ) .C1( )B(x1 x2 ) Bx2 (x1 x2 ) = . 2 C2 ( ) C1( ) = 1 + a11C1 ( ) + a21C2( ) . b1 C1( ) . a12C1( ) . x2C2( ) . 1 2 2 2 2 2C 2 ( 2 2 2 C2 ( ) ) C1 ( )C2( ) .x1 + x1C1( ) + x2 C2( ) + A ( ) . a21C2 ( ) 0 +1 2 +1 2 +1 2 We get three equations: 0 2 1C 2 ( 1 1 )+ 0 1 2 1C1( )C2( ) + 1 2 )C2( ) + 1 2 2 2 2 1C 2 ( 2 2 ) ) + x2 B (x1 x2 ) C2 ( ) . (a11x1 + a12 x2 + b1)C1( ) . b2C2( ) 2 1 2 C1 ( ) + 1 2 C1 ( )C2( ) + 1 2 ). 1 + C1 ( ) . This implies the initial conditions C1(0) = C2(0) = A(0) = 0: We want to find C1( ) C2( ) A( ) for > 0. 1 2 . 1 2 C1 ( ) . a11C1( ) . We have B (x1 x2 ) = . A ( ) B (x1 x2 ) Bx1 (x1 x2 ) = .C2( )B(x1 x2 ) 2 Bx1 x1 (x1 x2 ) = C1 ( )B(x1 x2 ) Bx1 x2 (x1 x2 ) = C1( )C2( )B (x1 x2 ) 2 Bx2 x2 (x1 x2 ) = C2 ( )B(x1 x2 ): 0 0 0 (IC) (PDE) becomes 0 = B (x1 x2 ) . 1 2 0 2 C 2( 1 1 1 1 2 1C1 ( )C2( ) . (a21x1 + a22x2 + b2)C2( ) 2 + 1 2 ( 1x1 + 2 x2 + )C1 ( ) + 1 2( 1x1 + 2x2 + )C1( )C2( ) 2 1 2 + 1 2 ( 1x1 + 2 x2 + )C2 ( ) 2 2 0 0 0 = x1 B (x1 x2 ) . 2 1 C2(0) = 0 2 A ( ) = b1C1( ) + b2C2( ) . 1 2 2C1 ( ) .

A( 0)g and the yield is . X2(t)C2( 0) .3 Calibration Let 0 > 0 be given.X1(t)C1( 0) . 32. A two-factor model (Duffie & Kan) 323 We first solve (1) and (2) simultaneously numerically. 1 2 . and then integrate (3) to obtain the function A( ).CHAPTER 32. The value at time t of a bond maturing at time t + 0 is B(X1 (t) X2(t) 0) = expf. Thus X2(t) = 1 X1(t)C1( 0 ) + X2(t)C2( 0) + A( 0)] : 0 This equation must hold for every value of X 1(t) and X2 (t). which implies that C1( 0) = 0 C2 ( 0) = We must choose the parameters 0 A( ) = 0: 1 2 a11 a12 b1 a21 a22 b2 so that these three equations are satisfied. 1 log B(X1(t) X2(t) 0) = 1 X1(t)C1( 0) + X2(t)C2( 0) + A( 0)] : 0 0 But we have set up the model so that X 2(t) is the yield at time t of a bond maturing at time t + 0 .

324 .

(t) (t) (t) d S(t) = S(t) (t) dW (t): The zero-coupon bond prices are given by ) # ( ZT B (t T ) = IE exp . Indeed. W is a Brownian motion relative to this filtration. but fF (t) 0 t T g This is not a geometric Brownian motion model. (0.1) fF (t) 0 t T g. For now. r(u) du F (t) " = IE 325 (t) F (t) (T ) t . We are particularly interested in the case that the interest rate is stochastic. we will have one asset. which we call a “stock” even though in applications it will usually be an interest rate dependent claim. The price of the stock is modeled by dS (t) = r(t) S (t) dt + (t)S (t) dW (t) where the interest rate process may be larger than the filtration generated by W . We shall work only under the risk-neutral measure.Chapter 33 Change of num´ raire e Consider a Brownian motion driven market model with time horizon T . which is reflected by the fact that the mean rate of return for the stock is r(t). We define the accumulation factor r(t) and the volatility process (t) are adapted to some filtration (t) = exp Zt 0 r(u) du so that the discounted stock price S (t) is a martingale. given by a term structure model we have not yet specified.

) F (t T ) = BSt(tT ) : ( Definition 33. Then IP N defined by IPN (A) = N1 (0) is risk-neutral for N . the stock We will say that a probability measure IP N is risk-neutral for the num´ raire N if every asset price. so 0 = IE (t) (S (T ) . e Theorem 0.2 (Bond as num´ raire) The T -maturity bond could be the num´ raire. e divided by N . The value of the forward contract at time t is zero. F (t T )B (t T ) Therefore. At time t. F (t T )B (t T ) (t) = S (t) . e e Example 33. T .71 Let N be a num´ raire. the e e S (t) B (t T ) stock is worth (t) units of money market and the T -maturity bond is worth (t) units of money market. We then denominate all other assets in units of this num´ raire. The T -forward price F (t T ) of the stock is the price set at time t for delivery of one share of stock at time T with payment at time T . The original probability measure IP is risk-neutral for the num´ raire (Example 33. Example 33.e. Z N (T ) dIP 8A 2 F (T ) (T ) A . the price process for some asset whose price is always e strictly positive.1 (Num´ raire) Any asset in the model whose price is always strictly positive can be e taken as the num´ raire. At time t e e is worth F(t T ) units of T -maturity bond and the T -maturity bond is worth 1 unit. is a martingale under IP N .1 (Money market as num´ raire) The money market could be the num´ raire.1). i. F (t T )) F (t) (T ) = (t)IE S (T ) Ft .326 so B(t T ) = IE 1 F (t) (t) (T ) is also a martingale (tower property). F (t T )IE (t) F (t) (T ) (T ) = (t) S (t) ..

Under IP . Therefore.. then N( IEN X F (t) = N (0)t)(t) IE N (0) T()T ) X F (t) N( T = N((tt)) IE N((T )) X F (t) : Therefore. i. and we see that IPN is a probability measure. Change of num´ raire e Note: 327 IP and IPN are equivalent. Y=N is a martingale. Y T Y IEN N(T ) F (t) = N((tt)) IE N((T )) N(T ) F (t) (T ) (T ) () = N((tt)) Y (tt) Y = N (t) (t) which is the martingale property for Y=N under IP N . For this. Let Y be an asset price.54). 33. and Z IP (A) = N (0) N((T )) dIPN 8A 2 F (T ): A T 1 Z N (T ) dIP IPN ( ) = N (0) (T ) 1 :IE N (T ) = N (0) (T ) (0) = N1 N(0) (0) =1 Proof: Because N is the price process for some asset. The risk-neutral measure for this num´ raire is e e 1 Z B (T T ) IPT (A) = B (0 T ) dIP A (T ) Z 1 = 1 B (0 T ) A (T ) dIP 8A 2 F (T ): . We must show that under IPN . N= is a martingale under IP .e. If 0 t T T and X is F (T )-measurable.1 Bond price as num´ raire e Fix T 2 (0 T ] and let B(t T ) be the num´ raire.CHAPTER 33. have the same probability zero sets. we need to recall how to combine conditional expectations with change of measure (Lemma 1. Y= is a martingale.

(1. we change the measure only “up to time T 1 using B (0 T ) B ((TT ) and only for A 2 F (T ). Denominated in units of T -maturity bond. We may assume without loss of generality that F (t T ) T g is a Brownian motion under T ) from now on.e.1) i. IPT T -forward measure..e.328 Because this bond is not defined after time T .1) is equal to the volatility F (t F (t T ) in (1. We may assume without loss of Theorem 2. and so has a differential of the form d F 1t) = (t T ) F 1t) dWS (t) ( ( where fWS (t) generality that (2.1) can be rewritten as T ) in (2. ) is called the stock is T ”.72 The volatility (t words. a differential without a dt term.1). the value of the ) F (t T ) = BSt(tT ) ( 0 t T: This is a martingale under IPT .1) 0 t T g is a Brownian motion under IPS . and so has a differential of the form dF (t T ) = We write F (t) rather than F (t F (t T )F (t T ) dWT (t) 0 t T t 0. In other d F 1t) = ( T ) F 1t) dWS (t) ( (2.2 Stock price as num´ raire e Let S (t) be the num´ raire.1’) .. the stock price is identically 1. 33. The riske e neutral measure under this num´ raire is e 1 IPS (A) = S (0) Z S (T ) (T ) dIP 8A 2 F (T ): A Denominated in shares of stock. (t T ) 0. (2. i. The process fWT 0 IPT . In terms of this num´ raire. the value of the T -maturity bond is B(t T ) = 1 : S (t) F (t) This is a martingale under IPS .

K )+ = IE S (T ) 1 S (T )>K .e. F (t T ) dWT (t) + 2 (t T ) dt F ( = F (t T ) F 1t) . KB (0 T )IPT fF (T ) > K g = S (0)IPS 1 < 1 .CHAPTER 33. Then 00 d F 1t) = dg (F (t)) ( = g (F (t)) dF (t) + 1 g (F (t)) dF (t) dF (t) 2 1 = . it F ( changes the mean return to zero.3 Merton option pricing formula The price at time zero of a European call is V (0) = IE (1T ) (S (T ) .. F 2 (t) F (t T )F (t T ) dWT (t) + F 31(t) 2 (t T )F 2(t T ) dt F 0 00 h i = F 1t) . Change of num´ raire e Proof: Let g (x) = 1=x.dWT (t) + F (t T ) dt]: ( Under IPT . KB (0 T )IPT fF (T ) > K g: F (T ) K f g f g f g f g .1) must be F (t T ) and WS must be WS (t) = . KIE 1 1 S (T )>K (T ) Z(T ) S (T ) dIP . (t T ) in (2. Under this measure. The change of measure from IPT to IPS makes F 1t) a martingale. Therefore.WT (t) + Zt 0 F (u T ) du: 33. so g 0(x) = . but the change of measure does not affect the volatility.WT is a Brownian motion. KB (0 T ) Z 1 = S (0) S (0) (T ) B(0 T ) (T ) dIP S (T )>K S (T )>K = S (0)IPS fS (T ) > K g . KB (0 T )IPT fS (T ) > K g = S (0)IPS fF (T ) > K g .1=x2 329 g (x) = 2=x3. i. F 1t) has volatility F (t T ) and mean ( rate of return 2 (t T ).

1 p T S (0) log KB (0 T ) . 1 F )T 2 1 p F T and we have the usual Black-Scholes formula. we have the following: T ). we still have the explicit V (0) = S (0)N ( 1) . 2 2 T F KB(0 T ) T T F = N ( 2) where 2= If r is constant.330 This is a completely general formula which permits computation as soon as we specify F (t we assume that F (t T ) is a constant F . 1 2 T : 2 F T) = e 1= rT . (0) log SK + (r + 1 2 )T 2 F F T S (0) 1 2 2 = p log K + (r . If 1 = B (0 T ) exp n W (T ) . 1 2 T F 2 (0 T IPT fF (T ) > K g = IPT F WT (T ) . 1 2 T > log KB (0 T ) 2 F S (0) T 1 p = IPT Wp(T ) > 1 log KB (0 T ) + 2 2 T F S (0) T F T 1 p p = IPT . then B (0 F . 1 2 T o F S 2 F F (T ) S (0) 1 S (0) IPS F (1T ) < K = IPS F WS (T ) .WT (T ) < 1 log S (0) . 1 2 T < log KB (0 T ) 2 F p S (0) S p = IPS Wp(T ) < 1 log KB (0 T ) + 1 F T 2 = N ( 1) 1 p T F T where 1= Similarly. KB(0 T )N ( 2): . When formula r is not constant. F T log S (0) 1 2 KB (0 T ) + 2 F T : o n F (T ) = BS (0) ) exp F WT (T ) .

We short bonds as necessary to finance this. KB(t T ) dN ( 2(t)): If X (0) = V (0). if F is constant. If X (t) is the value of the portfolio at time t. (t (t)S (t) will be invested in the bond. will X (t) = V (t) for 0 t T ? (3. Pk=11 i(t)Si(t) i k (t) = Sk (t) .t 1(t) = F N ( 1(t)) shares of stock and short We want to verify that this hedge is self-financing. the value of the portfolio will always be S (t)N ( 1(t)) . KN ( 2(t)) dB(t T ) . K dB(t T ) dN ( 2(t)) .KN ( 2(t))? If so. Will the position in the bond always be . KB (t T )N ( 2(t)) = V (t) and we will have a hedge.CHAPTER 33. the value of a European call expiring V (t) = S (t)N ( 1(t)) . KB (t T )N ( 2(t)) This formula also suggests a hedge: at each time t. then for 0 at time T is where 331 t T . At each time t.1) dV (t) = N ( 1(t)) dS (t) + S (t) dN ( 1(t)) + dS (t) dN ( 1(t)) . dX (t) = (t) dS (t) + X (B)(t T )(t) S (t) dB(t T ): The value of the option evolves according to (3. Mathematically. hold i (t) shares of asset i. t) 2 F F T ( 1 log FKt) .2) are difficult to compare. The number of shares of asset k held at time t is X (t) .2) Formulas (3. portfolio value evolves according to t . ( p1 . 1.73 Changes of num´ raire affect portfolio values in the way you would expect. so we simplify them by a change of num´ raire. . and invest the remaining wealth in asset k. e Proof: Suppose we have a model with k assets with prices S 1 S2 : : : Sk . t) : 2(t) = p 2 F T . i = 1 2 : : : k . t log FKt) + 1 2 (T . this question takes the following form. so the number of bonds owned is X B)(t T )(t) S (t) and the . Suppose we begin with $ V (0) and at each time t hold N ( 1(t)) shares of stock. Let (t) = N ( 1(t)): At time t. hold KN ( 2(t)) bonds. Theorem 3. Begin with a nonrandom initial wealth X (0). hold (t) shares of stock.1) and (3. then X (t) . 1 2 (T . Change of num´ raire e As this formula suggests. and let X (t) be the value of the portfolio at time t. e This change is justified by the following theorem.

iSi c Sk i=1 i=1 . . Pk=11 i Si i k= Sk X=N . Therefore. Pk=11 ici i =X c S: Sk . in advance.332 and X evolves according to the equation dX = = Note that k 1 X . . and define e 1 ::: i=1 k 1. . Pk=11 iSi=N i = Sk =N b . i=1 i Si ! dSk Sk Xk (t) = and we only get to specify Let N be a num´ raire. not k . . k c X c b k 1 c! dSk X b= dX i dSi + X . k X i=1 i=1 i dSi + i dSi: k X X. k 1 X . i (t)Si (t) X (t) b X (t) = N (t) Then c Si (t) = Si(t) N (t) i = 1 2 : : : k: 1 1 b 1 dX = N dX + X d N + dX d N ! i i i k k k 1 X dS + X S d 1 + X dS d 1 =N i i i i N i=1 i i N i=1 i=1 k X 1 dS + S d 1 + dS d 1 = = Now i=1 k X i=1 N N i N Si i dc: X .

We still hold (t) = N ( 1(t)) shares of stock at each time t.2)). i = 1 1. K dN ( 2(t)) = 0: This is a homework problem. the asset values are e Stock: Bond: The portfolio value evolves according to S (t) = F (t) B(t T ) B(t T ) = 1: B(t T ) (3. and all assets and the portfolio are denominated in units of N . K dN ( 2(t)): To show that the hedge works. the option value formula e V (t) = N ( 1(t))S (t) . We return to the European call hedging problem (comparison of (3. e In terms of the new num´ raire. but we now use the zero-coupon bond as num´ raire. we must show that (3. KN ( 2(t)) ( and b dV = N ( 1(t)) dF (t) + F (t) dN ( 1(t)) + dN ( 1(t)) dF (t) .CHAPTER 33.1) and (3. . (t)) d(1) = (t) dF (t): 1 In the new num´ raire. KB (t T )N ( 2(t)) becomes ) b V (t) = BVt(tT ) = N ( 1(t))F (t) . Change of num´ raire e This is the formula for the evolution of a portfolio which holds i shares of asset i. 333 2 : : : k.1’) b b dX (t) = (t) dF (t) + (X (t) .2’) F (t) dN ( 1(t)) + dN ( 1(t)) dF (t) .

334 .

(t T ) t ) | ({z T } B(t T ) dW (t): To implement HJM.1 Review of HJM under risk-neutral I P f (t T ) = df (t T ) = where The interest rate is r(t) = f (t Forward rate at time t for borrowing at time T: (t T ) (t T ) dt + (t T ) dW (t) (t T ) = ZT t (t u) du ( ZT ) # B (t T ) = IE exp . The bond prices " satisfy dB (t T ) = r(t) B(t T ) dt . t).Chapter 34 Brace-Gatarek-Musiela model 34. you specify a function 0 t T: A simple choice we would like to use is where > 0 is the constant “volatility of the forward rate”. r(u) du F (t) ( ZT t ) = exp . This is not possible because it leads to (t T ) = ( t T ) = f (t T ) ZT t f (t u) du df (t T ) = 2f (t T ) ZT t f (t u) du dt + f (t T ) dW (t) 335 ! . f (t u) du t volatility of T -maturity bond.

ct = 1=c. f (t) f (0) c c : f (t) = 1 . dt f (t) = 1 0 34. f (t) f (0) 0 1 = t .3) f (t v ) dv f (t t + u) du r(t u) du (2.r(t )D(t ): @ @T .2 Brace-Gatarek-Musiela model New variables: Current time t Time to maturity Forward rates: = T . (u = v . To see what problem this causes. t du = dv ) : = exp . Z t+ Zt Z0 0 (2. Jarrow and Morton show that solutions to this equation explode before T . t: (2.2) r(t ) = f (t t + ) r(t 0) = f (t t) = r(t) @ r(t ) = @ f (t t + ) @ @T Bond prices: D(t ) = B(t t + ) = exp .1) (2.4) @ D(t ) = @ B(t t + ) = . consider the similar deterministic ordinary differential equation f (t) = f 2(t) 0 where f (0) = c > 0. 1=c = ct .336 and Heath. We have This solution explodes at t 1 + 1 = Z t 1 du = t . f (t) = 1 f 2(t) d 1 . = exp . The problem with the above equation is that the dt term grows like the square of the forward rate. 1 = t . 1 .

6) @ @ (t ) = Z 0 (t u) du (2.1) = r(t 0) . 1 L(t 0) = : D(t )(1 + L(t 0)) = 1 ( ) .9) Also.10) 34. r(t )D(t ) dt (2.CHAPTER 34.6).5). $ D(t ) invested at time t in a (t + )-maturity bond grows to $ 1 at 0) is defined to be the corresponding rate of simple interest: Z@ r(t u) du 1 + L(t 0) = D(1 ) = exp t 0 nR o exp 0@ r(t u) du . L(t 1 = 4 year). t) rather than (t T ). In this notation.8) (t ) = (t ): ) and D(t ). dD(t ) = (2. (t )D(t ) dW (t): (2.7) (2.2) (2.8) differential applies only to first argument df (t {z+ ) | t } @ + @T f (t t + ) dt h i = @@ r(t ) + 1 ( (t ))2 dt + (t ) dW (t): 2 dB (t {z + ) | t } @ + @T B (t t + ) dt = (t ) (t ) dt + (t ) dW (t) + @@ r(t ) dt (2.6) We have We now derive the differentials of r(t dr(t ) = (2. Brace-Gatarek-Musiela model We will now write 337 (t ) = (t T .(2.(2.3 LIBOR Fix > 0 (say. r(t )] D(t ) dt . (t )B (t t + ) dW (t) . (t )B (t T ) dW (t) where (2. analogous to (2. time t + . the HJM model is df (t T ) = (t ) (t ) dt + (t ) dW (t) dB (t T ) = r(t)B(t T ) dt .4) differential applies only to first argument = r(t) B (t t + ) dt .5) (2.5) and (2.

0 + r(t u) du = exp .2) o r(t u) du .338 34. The forward LIBOR L(t ) is defined to be the simple (forward) interest rate for this investment: D(t + ) (1 + L(t )) = 1 D(t ) R D(t ) = exp f. Can do this at time t by shorting DDt(t +) ) bonds maturing at time t + and going long one bond maturing at time t + + .1. 1 o r(t ) is the continuously compounded rate.5) so that dL(t ) = (: : : ) dt + L(t ) (t ) dW (t) . For fixed positive .5 The dynamics of L(t ) We want to choose (t ) t 0 0. We cannot have a log-normal model for r(t ) because solutions explode as we saw in Section 34. appearing in (2. 1 >0 fixed: (4.4 Forward LIBOR ( At time t. with payback of $1 at time ( t + + . D(t (t +) ) D(t ) + D(t + ) = 0: D (Z nR + r(t u) du ) L(t ) = Connection with forward rates: exp + r(t u) du . 34. L(t ) is the simple rate over a period of duration .1) ( @ exp Z @ so + r(t u) du ) =0 = r(t + ) exp = r(t ) (Z + r(t u) du ) =0 f (t t + ) = r(t ) = lim L(t ) = exp nR + #0 exp nR + r(t u) du . agree to invest $ DDt(t +) ) at time t + . 0 r(t u) dugo n R 1 + L(t ) = D(t + ) exp . 1 o : (4. The value of this portfolio at time t is > 0 is still fixed. we can have a log-normal model for L(t ).

(5. 1 5 dL(t ) (4:1) d 4 = (Z + ) Z + 1 = exp ( ) 1 exp Z + r(t u) du d Z +2 (4. Recall (2. (t )]2 dt 2 . r(t ) + 2 ( (t + ))2 .9): i @ h dr(t ) = @u r(t u) + 1 ( (t u))2 dt + (t u) dW (t): 2 Therefore. (t )] dW (t) : 1 r(t + ) . and is a subclass of HJM models.1) and o 3 2 nR + exp r(t u) du . 1 ( (t ))2] dt 2 + (t + ) . (t )] dW (t) dr(t u) du (5. r(t ) + 2 ( (t + ))2 .1). (t )] dW (t) + 1 (t + ) .CHAPTER 34. Brace-Gatarek-Musiela model 339 0 0. 1 ( (t ))2 dt 2 + (t + ) . for some (t ) t corresponding to particular choices of (t ).2) = 1 1 + L(t )] r(t + ) .1) 1 = 1 + L(t )] r(t u) du d r(t u) du + r(t u) du !2 (5. d Z + r(t u) du = = ! Z Z + + @ hr(t u) + 1 ( (t u))2i du dt + Z + (t u) du dW (t) 2 i h @u 1 = r(t + ) . (t )] dt = + (t + ) . r(t )] dt + (t + ) (t + ) . This is the BGM model.

340 But 2 nR @ L(t ) = @ 4 exp @ @ (Z + = exp + r(t u) du . r(t )]: Therefore.6 Implementation of BGM Obtain the initial forward LIBOR curve L(0 ) (t ) 0 0: from market data. (t )]: (t + ) dt + dW (t)]: Take (t ) to be given by (t )L(t ) = 1 1 + L(t )] (t + ) .4) ) (t + ) = (t ) + L(t L(t(t ) ) : 1+ Plugging this into (5. (t )]: (5. 1 5 o 3 r(t u) du : r(t + ) .3) Then dL(t ) = @@ L(t ) + (t )L(t ) (t + )] dt + (t )L(t ) dW (t): Note that (5. dL(t ) = @@ L(t ) dt + 1 1 + L(t )] (t + ) .4) yields (5.3) is equivalent to (5. Choose a forward LIBOR volatility function (usually nonrandom) t 0 .4’) 34.3’) 2 ) 2 dL(t ) = @@ L(t ) + (t )L(t ) (t ) + L1 (t L(t (t ) ) dt + + (t )L(t ) dW (t): " # (5. r(t )] ) = 1 1 + L(t )] r(t + ) .

the resulting Remark 34. (t ) is usually taken to be nonrandom. then @ (t )L(t )! @ (t + ) =0 = (t ) and so We saw before (eq. t): #0.3’). we obtain (t ) for L(t ) t 0 and we continue recursively.4’) to obtain L(t ) t 0 0 Plugging the solution into (5. t) dt dK (t T ) = (t T . 4. This is not as terrible as it first appears.4’)) is a stochastic partial differential equation because of the @ @ L(t ) term. (t ) : If we let #0.2) that as (t T . t) dt + dW (t) : 1 + K (t T ) (6.CHAPTER 34.3 From (5. we must also choose a partial bond volatility function (t ) With these functions.3’). 2 0 ) solve (5. t) .4) (equivalently.4’) to obtain < : <2 <2 ) generated recursively by (5. t)K (t T )! (t T . set K (t T ) = L(t T . (5. @@ L(t T . t) dt + K (t T ) (t T . t + ) dt + dW (t)] (t T . t): Then dK (t T ) = dL(t T . (t ) is random. t)K (t T ) (t T .2 (5. Brace-Gatarek-Musiela model 341 Because LIBOR gives no rate information on time periods smaller than .4) and (5.1 BGM is a special case of HJM with HJM’s (t In BGM.4’) become Remark 34. Returning to the HJM variables t and T . Remark 34. L(t )!r(t ) = f (t t + ) .3) we have (t )L(t ) = 1 + L(t )] (t + ) .1) and (5. t)K (t T ) = (t T . . we can for each t 0 0 < for maturities less than from the current time variable t. We then solve (5.

solutions 1+K (t T ) . and we can use it to switch to the forward measure Girsanov’s Theorem implies that WT (t) = W (t) + is a Brownian motion under IPT . Z t (u T . Zt 0 (u T .1) is given by equation (2. t) (t T . u) du 0 t T .6) we have o d B(t(tT ) = 1t) . 2(t T . t) dW (t): (t) The solution B (ttT ) to this stochastic differential equation is given by () B (t T ) = exp . B (t T ) (t T . t)K 2(t T ) K (t T ) 2(t T .5): df (t T ) = (t T . t)K 2 (t T ) 1 + K (t T ) T . u) dW (u) . u))2 du : 2 0 (t)B (0 T ) 0 1 Z 1 dIP IPT (A) = B (0 T ) Z B(TAT )(T ) = dIP 8A 2 F (T ): A (T )B (0 T ) This is a martingale. 1 Z t( (u T .4 Although the dt term in (6. t) dt + dW (t)] : 2 Remark 34.r(t)B(t T ) dt + dB(t T )] ) ( = .7 Bond prices Let (t) = exp nR t 0 r(u) du : From (2. t)K (t T ): 2(t 34.342 so K (t T )!f (t T ): Therefore.1) has the term to this equation do not explode because (t T t)K 2 (t T ) involving K 2 . the limit as #0 of (6.

t)K (t T ) (t T .9 Pricing an interest rate caplet Consider a floating rate interest payment settled in arrears. u) du ) (8. 34. Brace-Gatarek-Musiela model 343 34. 1 2 Zt 0 2(u T . 2 2(t). t + ) dt + dW (t)] = (t T . Thus. c)+ F (t) t) = (K (T T ) . At time T + . u) dWT + (u) .1) (u T . u) dWT + (u) . the value of the caplet at time T + is (K (T T ) .CHAPTER 34.8 Forward LIBOR under more forward measure From (6. We determine its value at times 0 t T + . A caplet protects its owner by requiring him to pay only the cap c if K (T T ) > c. u) dWT + (u) . c)+ B (t T + ): (9. 1 2 2(t) = ZT t 2(u T . Then X (t) = is normal with variance ZT t (u T . c)+ . u) du K (T T ) = K (0 T ) exp = K (t T ) exp We assume that (Z T 0 (Z T t (u T . c)+ IE (T (+ ) F (t) = (K (T T ) . u) du 1 and mean . 1 2 T . t) CT + (t) = IE (T(+ ) (K (T T ) . t)K (t T ) dWT + (t) so K (t T ) = K (0 T ) exp and Zt 0 (u T . u) du : ) is nonrandom.2) ZT t 2(u T .1) we have dK (t T ) = (t T . u) du (8. u) dWT + (u) .1) . ZT 2(u 1 2 0 ZT t 2 (u T . the floating rate interest payment due is L(T 0) = K (T T ) the LIBOR at time T . Case I: T t T+ .

1) and (8. c)+ F (t) : Set g(y) = IET + (y expfX (t)g .t and (9. IPT + (A) = where Z A Z (T + ) dIP 8A 2 F (T + ) T Z (t) = (B (t (0 + +) ) : t)B T We have t) CT + (t) = IE (T(+ ) (K (T T ) . c)+ F (t) From (8. c N (t) c Then h i 1 log y . c)+ 1 = y N 1 log y + 2 (t) . we have (t) = p T . Furthermore. c)+ F (t)7 7 6 = B (t T + ) B (t T + ) 7 6 (T + )B(0 T + ) 5 4| {z } {z } | 1 Z (t) Z (T + ) = B (t T + )IET + (K (T T ) . : In the case of constant . c)+ F (t) 2 3 7 6 (t)B (0 T + ) IE 6 B (T + T + ) (K (T T ) .344 Case II: 0 Recall that t T. .2) is called the Black caplet formula. u) du and mean .2) we have K (T T ) = K (t T ) expfX (t)g R 1 where X (t) is normal under IPT + with variance 2(t) = tT 2(u T . X (t) is independent of F (t). CT + (t) = B(t T + )IET + (K (t T ) expfX (t)g .2) CT + (t) = B(t T + ) g (K (t T )) 0 t T . 2 2(t). 1 (t) : ( t) c 2 (9.

we may assume that is constant on each of the intervals . 1: The value at time t of the cap is the value of all remaining caplets. (t ) = ( ): Then g depends on ZT 0 If we know the caplet price CT + know caplet prices where T0 R (0).e.1) (0 T0) (T0 T1) : : : (Tn 1 Tn ) . C (t) = 34. we can “back out” CT0+ (0) CT1+ (0) : : : CTn+ (0) 2(v ) dv .10 Pricing an interest rate cap Let T0 = 0 T1 = T2 = 2 : : : Tn = n : (K (Tk Tk ) .11 Calibration of BGM The interest rate caplet c on L(0 X k:t Tk CTk (t): T ) at time T + has time-zero value CT + (0) = B(0 T + ) g (K (0 T )) where g (defined in the last section) depends on ZT 0 2 (u T .. Z Tn Tn. u) du = ZT 0 2(v ) dv: Z T0 0 < T1 < : : : < Tn . Brace-Gatarek-Musiela model 345 34. 2(v ) dv Z T1 T0 2(v ) dv = Z T0 0 2(v ) dv ::: In this case. i..CHAPTER 34. u) du: Let us suppose is a deterministic function of its second argument. we can “back out” the squared volatility 0T 2(v ) dv . i. If we Z T1 0 2(T .1 2(v ) dv: (11. c)+ at time Tk+1 A cap is a series of payments k = 0 1 : : : n .e.

Let n be a large fixed positive integer.6)). and To implement BGM.1). )= 0. The long rate is n 1 log 1 = 1 X log 1 + L(t (k . so that n is 20 or 30 years. Then D(t n ) = exp = = n Y k=1 n Y k=1 1 (Z n 0 exp (Z k r(t u) du (k 1) . we can “back out” 0T 2(v ) dv and then differenp 2( ) for all 0. and set T1 = T0 + T2 = T0 + 2 : : : Tn = T0 + n : . and 0 4 (t ) = 0 We can now solve (or simulate) to get t 0 0 0 < : L(t ) t 0 or equivalently.6. K (t T ) t 0 T 0 using the recursive procedure outlined at the start of Section 34.12 Long rates The long rate is determined by long maturity bond prices. 34. t 0 0 < : t+ (see (2. it is reasonable to set Since is small (say 1 year). 1) )] where the last equality follows from (4. Now (t ) is the volatility at time t of a zero coupon bond maturing at time < . 1) )]: n D(t n ) n k=1 34. we need both ( ) (t ) and choose these constants to make the above integrals have the values implied by the caplet prices. ) r(t u) du ) 1 + L(t (k .346 If we know caplet prices CT + tiate to discover 2( ) and ( R (0) for all T 0.13 Pricing a swap Let T0 0 be given.

the value of the swap is n 1 X . the swap formula is generic. which depends on the term structure model and requires estimation of . c) F (t) k+1 = IE (T(t) ) B (T 1T ) . . (1 + c)B(t Tk+1)] value of the swap equal to zero: Bt ) ( wT0 (t) = B (t (T T0+ . c) F (t) : k+1 k=0 Now 1 + L(Tk 0) = B (T 1T ) k k+1 so L(Tk 0) = 1 B(T 1T ) . c F (t) We compute k k+1 3 2 k+1 6 7 6 (Tk ) F (T ) F (t)7 . B (t Tn ): The forward swap rate wT0 (t) at time t for maturity T0 is the value of c which makes the time-t . 1 . cB (t Tn ) .: B+tBT(n )T )] : t n 1) : : In contrast to the cap formula. (1 + c)B (t T2) + : : : + B (t Tn 1 ) . c) F (t) k+1 k=0 = n 1 X . cB (t T2) . : : : . 1 : k k+1 IE (T(t) ) (L(Tk 0) . c) at time Tk+1 k = 0 1 : : : n . k=0 B (t Tk ) . (1 + c)B (t Tk+1): = IE B (Tk Tk+1 ) n 1 X . (1 + c)B (t Tn ) = B (t T0) . (1 + c)B (t T1) + B (t T1) . cB (t T1) . (L(Tk 0) . (1 + c)B (t Tk+1) = B (t Tk ) . 1: IE (T(t) ) (L(Tk 0) . = B (t T0) . Brace-Gatarek-Musiela model The swap is the series of payments For 0 347 t T0. (1 + c)B (t T ) (t) 7 6 = IE 6 (T )B (T T ) IE (T ) k k+1 7 k+1 5 4 k k k+1 | {z } The value of the swap at time t is (t) (Tk+1 ) F (t) .CHAPTER 34. IE (T(t) ) (L(Tk 0) .

348 .

and then binomial versions are developed for purposes of implementation. including Cox & Rubinstein (1985).Chapter 35 Notes and References 35. Accounts of this model can be found in several places. although the concept dates back to 1934 work of L´ vy. These can be used to compute expectations and variances. In the first stage. This point of view handles both discrete and continuous random variables within the same unifying framework. A conditional expectation is itself a random variable. 349 . In this stage one views a random variable as a function from a sample space to the set of real numbers IR. The second stage of probability theory is measure theoretic. A well-written book on measure-theoretic probability is Billingsley (1986). measurable with respect to the conditioning -algebra. and even conditional expectations. Probability theory is usually learned in two stages. The binomial asset pricing model was developed by Cox. A succinct book on measure-theoretic probability and martingales in discrete time is Williams (1991). one learns how transformations of continuous random variables cause changes in their densities. A more detailed book is Chung (1968). Each set A in F has a probability IP (A). Ross & Rubinstein (1979). The term martingale was apparently first used by Ville (1939). Certain subsets of are called events. e 35. Many models are first developed and understood in continuous time. and the collection of all events forms a -algebra F . Dothan (1990) and Ritchken (1987). Furthermore. one learns that a discrete random variable has a probability mass function and a continuous random variable has a density. A well-written book which contains all these things is DeGroot (1986). This point of view is indispensible for treating the rather complicated conditional expectations which arise in martingale theory. The first complete account of martingale theory is Doob (1953).2 Binomial asset pricing model.1 Probability theory and martingales. The measure-theoretic view of probability theory was begun by Kolmogorov (1933).

The first mathematically rigorous construction of Brownian motion was carried out by Wiener (1923. but they do not provide information which permits an investor to outperform the market. This motion is now known to be caused by the buffeting of the pollen by water molecules. An excellent reference for practitioners. with a minimum of fuss. even though Brownian motion can take negative values. The integral with respect to Brownian motion was developed by Itˆ (1944).4 Stochastic integrals. L´ vy (1939. Some empirical studies supporting the efficient market hypothesis are Kendall (1953). Osborne (1959). Boness (1964). including Billingsley (1986). and Jarrow (1988). Huang & Litzenberger (1988). A mathematical construction of this integral. include sections on Itˆ ’s o integral and formula. Samuelson (1965. It was introduced to o finance by Merton (1969). Chung & Williams (1983). Stochastic calculus begins with Itˆ (1944). 35. 1973) presents the argument that geometric Brownian motion is a good model for stock prices. 35. Some mathematics texts on stochastic calculus are Øksendal (1995). Dothan (1990) and Duffie (1992). a much weaker condition than the claim that stock prices follow a geometric Brownian motion.350 35. past stock prices may be useful to estimate the parameters of the distribution of future returns. which asserts that all information which can be learned from technical analysis of stock prices is already reflected in those prices. Sprenkle (1961). The mathematical formulation of the efficient market hypothesis is that there is a probability measure under which all discounted stock prices are martingales. 1924). The development in this course is a summary of that found in Karatzas & Shreve (1991). including (in order of increasing o mathematical difficulty) Hull (1993). According to this hypothesis. In 1828 Robert Brown observed irregular movement of pollen suspended in water. A criticism of the efficient market hypothesis is provided by LeRoy (1989). The last of these papers discusses other distributions which fit stock prices better than geometric Brownian motion. Brownian motion and its properties are presented in a numerous texts. now in preprint form. 1948) discove ered many of the nonintuitive properties of Brownian motion. Bachelier (1900) used Brownian motion (not geometric Brownian motion) as a model of stock prices. Ingersoll (1987). A provocative article on the source of . The quadratic variation of martingales was introduced by Fisk (1966) and developed into the form used in this course by Kunita & Watanabe (1967).5 Stochastic calculus and financial markets. Alexander (1961) and Fama (1965). is Musiela & Rutkowski (1996). is given by Øksendal (1995).3 Brownian motion. as explained by Einstein (1905). Some other books on dynamic models in finance are Cox & Rubinstein (1985). Protter (1990) and Karatzas & Shreve (1991). This is often confused with the efficient market hypothesis. Many finance books.

although the result for constant was established much earlier by Cameron & Martin (1944).) Even though geometric Brownian motion is a less than perfect model for stock prices. including discussions of the Kolmogorov forward and backward equations. or even through dependence on the paths of other assets.6 Markov processes. and risk-neutral measures. (This and many other papers by Merton are collected in Merton (1990).7 Girsanov’s theorem. see Karatzas & P Shreve (1991). the Brownian-motion driven model is mathematically the most general possible for asset prices without jumps. Chung & Williams (1983). When both the interest rate and volatility of an asset are allowed to be stochastic. 35. which makes good reading even today. connecting stochastic differential equations to partial differential equations. Another alternative model for the stock price underlying options. A numerical treatment of the partial differential equations arising in finance is contained in Wilmott. Øksendal (1995). The form of the martingale representation theorem presented here is from Kunita & Watanabe (1967). This methodology frees the Brownian-motion driven model from the assumption of constant interest rate and volatility. The theorem requires a f technical condition to ensure that IEZ (T ) = 1. the Black-Scholes option hedging formula seems not to be very sensitive to deficiencies in the model. . Markov processes which are solutions to stochastic differential equations are called diffusion processes. page 182. these parameters can be random through dependence on the path of the underlying asset. so that I is a probability measure. due to F¨ llmer & Schweizer (1993). It can also be found in Karatzas & Shreve (1991). Notes and References stock price movements is Black (1986). Kloeden & Platen (1992) is a thorough study of the numerical solution of stochastic differential equations. all treat this subject. Protter (1990). and Karatzas & Shreve (1991). is due to Feyman (1948) and Kac (1951). is Chapter 15 of Karlin & Taylor (1981). The other books cited previously. The application of the Girsanov Theorem and the martingale representation theorem to risk-neutral pricing is due to Harrison & Pliska (1981). An important companion paper is Merton (1973). A good introduction to this topic. page 198. 1995) and also Duffie (1992). the martingale representation theorem. o is similar to that originally given by Black & Scholes (1973). using only Itˆ ’s formula. Dewynne and Howison (1993. has o the geometric Ornstein-Uhlenbeck process as a special case. 351 The first derivation of the Black-Scholes formula given in this course. The constant elasticity of variance model for option pricing appears in Cox & Ross (1976). Girsanov’s Theorem in the generality stated here is due to Girsanov (1960). 35.CHAPTER 35. The Feynman-Kac Theorem.

M¨ ller (1989) and Taqqu & Willinger (1987). that the absence of arbitrage implies the existence of a risk-neutral measure.352 When asset processes have jumps. as stated here. 1995). We do not provide an explicit pricing formula here. Delbaen (1992). Back (1991). Sosin & Shepp (1979) and Conz´ & Viswanathan (1991). including Stricker (1990).9 American options. Lookback options have been studied by Goldman. Merton (1976). The reflection principle. Naik & Lee (1990). Shirakawa (1990. Goldman. Jones (1984). In addition to the fundamental papers of Harrison & Kreps (1979). Delbaen & Schachermayer (1994a. Beinert & Trautman (1991). Some works on hedging and/or optimization in models which allow for jumps are Aase (1993). but in continuous-time models. For the continuous-time case. and Harrison & Pliska (1981. risk-free hedging is generally not possible. pages 196–197. some other works on the relationship between market completeness and uniqueness of the risk-neutral measure are Artzner & Heath (1990). Bryis & Crouhy (1994) provide an approximate pricing formula. (1990) for a treatment of the American option optimal stopping problem via variational inequalities. The perpetual American put problem was solved by McKean (1965). i. It is tempting to believe the converse of Part I. which leads to a justification of the Brennan-Schwartz algorithm). by Cox.. 1983). Lakner (1993). Bates (1988. Delbaen (1992).c). Schweizer (1992a. and Geman & Yor (1993) obtain the Laplace transform of the price. Approximation and/or numerical solutions for the American option problem have been proposed by several authors. Jacka (1992).e. Elliott & Kopp (1990).b). although the partial differential equation given here by the Feynman-Kac Theorem characterizes the exact price. a slightly stronger condition is needed to guarantee existence of a risk-neutral measure. Sosin & Gatto (1979). Bouaziz. Madan & Milne (1991). Madan & Seneta (1990). 1983). is taken from Karatzas & Shreve (1991). Rogers & Shi (1995) provide a lower bound.b).1991) and Xue (1992). This is true in discrete-time models. adjusted to account for drift. results have been obtained by many authors. e Because it is difficult to obtain explicit formulas for the prices of Asian options. Mercurio & Runggaldier (1993).1992). Brennan & Schwartz (1977) (see Jaillet et al. Explicit formulas for the prices of barrier options have been obtained by Rubinstein & Reiner (1991) and Kunitomo & Ikeda (1992). u 35. most work has been devoted to approximations. and Fritelli & Lakner (1994. A general arbitrage-based theory for the pricing of American contingent claims and options begins with the articles of Bensoussan (1984) and Karatzas (1988). Jarrow & Madan (1991a). Ross & Rubinstein (1979) (see Lamberton (1993) for the convergence of the associated binomial and/or finite difference schemes) . Jarrow & Madan (1991b. can be found in Harrison & Pliska (1981.8 Exotic options. The Fundamental Theorem of Asset Pricing. see Myneni (1992) for a survey and additional references. 35. including Black (1975).

Notes and References 353 and by Parkinson (1977). Our presentation of this material is similar to that of Duffie & Stanton (1992). . Bunch & Johnson (1992).b). and Longstaff & Schwartz (1992a. and a discrete-time version is provided by Heath.10 Forward and futures contracts. The distinction between futures contracts and daily resettled forward contracts has only recently been recognized (see Margrabe (1976).12 Change of num´ raire. Black (1976)) and even more recently understood. The prices for differential swaps have been worked out by Jamshidian (1993a. Barone-Adesi & Whalley (1987). MacMillan (1986). 35. e This material in this course is taken from Geman. Foreign exchange options were priced by Biger & Hull (1983) and Garman & Kohlhagen (1983). one may consult Duffie (1989). Barone-Adesi & Elliott (1991). Derman & Toy (1990). Jarrow & Morton (1990). 35. The Merton option pricing formula appears in Merton (1973). The Brace-Gatarek-Musiela variation of the HJM model is taken from Brace. equilibrium setting. 1993b) and Brace & Musiela (1994a). which also considers options on futures. and to Chapte 7 of Duffie (1992). A summary of this model appears as Reed (1995). Johnson (1983). Geske & Johnson (1984). and Carr & Faguet (1994). Ho & Lee (1986). The Cox-Ingersoll-Ross model is presented in (1985a. et al.CHAPTER 35.b). Ingersoll & Ross (1981) and Jarrow & Oldfield (1981) provide a discrete-time arbitrage-based analysis of the relationship between forwards and futures. Brennan & Schwartz (1979. 1982) (but see Hogan (1993) for discussion of a problem with this model). Related works on term structure models and swaps are Flesaker & Hughston (1995) and Jamshidian (1996).11 Term structure models. For additional reading on forward and futures contracts. A partial list of other term structure models is Black. Duffie & Kan (1993). Broadie & Detemple (1994). The Hull & White (1990) model is a generalization of the constant-coefficent Vasicek (1977) model. The continuous-time Heath-Jarrow-Morton model appears in Heath.b). (1995). El Karoui and Rochet (1995). Other surveys of term structure models are Duffie & Kan (1994) and Vetzal (1994). Jarrow & Morton (1992).13 Foreign exchange models. Implementations of the model appear in Hull & White (1994a. Jamshidian (1990).b). The presentations of these given models here is taken from Rogers (1995). Carverhill & Pang (1995) discuss implementation. Omberg (1987). Cox. Similar ideas were used by by Jamshidian (1989). whereas Richard & Sundaresan (1981) study these claims in a continuous-time. Brace & Musiela (1994a. 35. 35.

354

35.14 REFERENCES

(1993) Aase, K. K., A jump/diffusion consumption-based capital asset pricing model and the equity premium puzzle, Math. Finance 3(2), 65–84. (1961) Alexander, S. S., Price movements in speculative markets: trends or random walks, Industrial Management Review 2, 7–26. Reprinted in Cootner (1964), 199–218. (1990) Artzner, P. & Heath, D., Completeness and non-unique pricing, preprint, School of Operations Research and Industrial Engineering, Cornell University. ´ (1900) Bachelier, L., Th´ orie de la sp´ culation, Ann. Sci. Ecole Norm. Sup. 17, 21–86. Reprinted e e in Cootner (1964). (1991) Back, K., Asset pricing for general processes, J. Math. Econonics 20, 371–395. (1991) Barone-Adesi, G. & Elliott, R., Approximations for the values of American options, Stoch. Anal. Appl. 9, 115–131. (1987) Barone-Adesi, G. & Whalley, R., Efficient analytic approximation of American option values, J. Finance 42, 301–320. (1988) Bates, D. S., Pricing options under jump-diffusion processes, preprint, Wharton School of Business, University of Pennsylvania. (1992) Bates, D. S., Jumps and stochastic volatility: exchange rate processes implicit in foreign currency options, preprint, Wharton School of Business, University of Pennsylvania. (1991) Beinert, M. & Trautmann, S., Jump-diffusion models of stock returns: a statistical investigation, Statistische Hefte 32, 269–280. (1984) Bensoussan, A., On the theory of option pricing, Acta Appl. Math. 2, 139–158. (1983) Biger, H. & Hull, J., The valuation of currency options, Financial Management, Spring, 24–28. (1986) Billingsley, P., Probability and Measure, 2nd edition, Wiley, New York. (1975) Black, F., Fact and fantasy in the use of options, Financial Analyst’s J., 31, 36–41, 61–72.

CHAPTER 35. Notes and References
(1976) Black, F., The pricing of commodity contracts, J. Financial Economics 3. (1986) Black, F., Noise, J. Finance 41, 529–543.

355

(1990) Black, F., Derman, E. & Toy, W., A one-factor model of interest rates and its application to treasury bond options, Fin. Anal. Journal, 33–39. (1973) Black, F. & Scholes, M., The pricing of options and corporate liabilities, J. Political Economy 81, 637–659. (1964) Boness, Some evidence of the profitability of trading put and call options. In Cootner (1964), 475–496. (1994) Bouaziz, L., Bryis, E. & Crouhy, M., The pricing of forward-starting Asian options, J. Banking Finance. (1994a) Brace, A. & Musiela, M., Swap derivatives in a Gaussian HJM framework, preprint, School of Mathematics, University of New South Wales. (1994b) Brace, A. & Musiela, M., A multi-factor Gauss Markov implementation of Heath, Jarrow and Morton, Math. Finance 4, 259–283. (1995) Brace, A.. Gatarek, D. & Musiela, M., The market model of interest rate dynamics, preprint, School of Mathematics, University of New South Wales. (1977) Brennan, M. J. & Schwartz, E. S., The valuation of the American put option, J. Finance 32, 449–462. (1979) Brennan, M. J. & Schwartz, E. S., A continuous-time approach to the pricing of bonds, J. Banking Finance 3, 133–155. (1982) Brennan, M. J. & Schwartz, E. S., An equilibrium model of bond pricing and a test of market efficiency, J. Financial Quantitative Analysis 17, 301–329. (1994) Broadie, M. & Detemple, J., American option values: new bounds, approximations and a comparison of existing methods, preprint, Graduate School of Business, Columbia University. (1992) Bunch, D. & Johnson, H., A simple and numerically efficient valuation method for American puts, using a modified Geske-Johnson approach, J. Finance 47, 809–816. (1944) Cameron, R. H. & Martin, W. T., Transformation of Wiener integrals under translations,

356 Ann. Math. 45, 386–396. (1994) Carr, P. & Faguet, D., Fast accurate valuation of American options, preprint, Johnson School of Business, Cornell University. (1995) Carverhill, A. & Pang, K., Efficient and flexible bond option valuation in the Heath, Jarrow, and Morton framework, J. Fixed Income, September, 70–77. (1968) Chung, K. L. A Course in Probability Theory, Academic Press, Orlando, FL. (1983) Chung, K. L. & Williams, R. J. (1983) Introduction to Stochastic Integration, Birkh¨ user, a Boston. (1991) Conz´ , A. & Viswanathan, R., Path dependent options: the case of lookback options, J. e Finance 46, 1893–1907. (1964) Cootner, P. H. (editor), The Random Character of Stock Market Prices, MIT Press, Cambridge, MA. (1981) Cox, J. C., Ingersoll, J. E., & Ross, S, The relation between forward prices and futures prices, J. Financial Economics 9, 321–346. (1985a) Cox, J. C., Ingersoll, J. E. & Ross, S., An intertemporal general equilibrium model of asset prices, Econometrica 53, 363–384. (1985b) Cox, J. C., Ingersoll, J. E. & Ross, S., A theory of the term structure of interest rates, Econometrica 53, 373–384. (1976) Cox, J. C. & Ross, S. A., The valuation of options for alternative stochastic processes, J. Fin. Economics 3, 145–166. (1979) Cox, J. C., Ross, S. A. & Rubinstein, M., Option pricing: a simplified approach, J. Financial Economics 7, 229–263. (1985) Cox, J. C. & Rubinstein, M. Options Markets, Prentice Hall, Englewood Cliffs, NJ. (1986) DeGroot, M., Probability and Statistics, 2nd edition, Addison Wesley, Reading, MA. (1992) Delbaen, F., Representing martingale measures when asset prices are continuous and bounded, Math. Finance 2(2), 107–130.

CHAPTER 35. Notes and References

357

(1994a) Delbaen, F. & Schachermayer, W., A general version of the fundamental theorem of asset pricing, Mathematische Annalen 300, 463–520. (1994b) Delbaen, F. & Schachermayer, W., Arbitrage and free-lunch with bounded risk, for unbounded continuous processes, Math. Finance 4(4), 343–348. (1953) Doob, J., Stochastic Processes, Wiley, New York. (1990) Dothan, M. U., Prices in Financial Markets, Oxford University Press, New York. (1989) Duffie, D., Futures Markets, Prentice-Hall, Englewood Cliffs, NJ. (1992) Duffie, D., Dynamic Asset Pricing Theory, Princeton University Press, Princeton. (1993) Duffie, D. & Kan, R., A yield-factor model of interest rates, preprint. (1994) Duffie, D. & Kan, R., Multi-factor term structure models, Phil. Transactions Royal Society London A 347, 577–586. (1992) Duffie, D. & Stanton, R., Pricing continuously resettled contingent claims, J. Economic Dynamics and Control, 16, 561–573. (1905) Einstein, A. On the movement of small particles suspended in a stationary liquid demanded by the molecular-kinetic theory of heat, Ann. Physik 17. (1990) Elliott, R. J. & Kopp, P. E., Option pricing and hedge portfolios for Poisson processes, Stoch. Anal. Appl. 9, 429–444. (1965) Fama, E. The behavior of stock-market prices, J. Business 38, 34–104. (1948) Feynman, R., Space-time approach to nonrelativistic quantum mechanics, Rev. Mod. Phys. 20, 367–387. (1966) Fisk, D. L., Sample quadratic variation of continuous, second-order martingales, Z. Wahrscheinlichkeitstheorie verw. Gebiete 6, 273–278. (1995) Flesaker, B. & Hughston, L., Positive interest, preprint. (1993) F¨ llmer, H. & Schweizer, M., A microeconomic approach to diffusion models for stock o prices, Math. Finance 3, 1–23.

358 (1994) Fritelli, M. & Lakner, P., Almost sure characterization of martingales, Stochastics 49, 181– 190. (1995) Fritelli, M. & Lakner, P., Arbitrage and free-lunch in a general finance market model: the fundamental theorem of asset pricing, IMA Vol 65: Mathematical Finance, 89–92, (M. Davis, D. Duffie, W. Fleming & S. Shreve, eds.), Springer-Verlag, New York. (1983) Garman, M. & Kohlhagen, S., Foreign currency option values, J. International Money Finance 2, 231–237. (1995) Geman, H., El Karoui, N. & Rochet, J.-C., Changes of num´ raire, changes of probability e measure, and option pricing, J. Applied Probability 32, 443–458. (1993) Geman, H. & Yor, M., Bessel processes, Asian options, and perpetuities, Mathematical Finance 3, 349–375. (1984) Geske, R. & Johnson, H., The American put-option valued analytically, J. Finance 39, 1511– 1524. (1960) Girsanov, I. V., On transforming a certain class of stochastic processes by absolutely continuous substitution of measures, Theory Probability Applications 5, 285–301. (1979) Goldman, B., Sosin, H. & Gatto, M., Path dependent options: “Buy at the low, sell at the high,” J. Finance 34, 69–84. (1979) Goldman, B., Sosin, H. & Shepp, L. A., On contingent claims that insure ex-post optimal stock market timing, J. Finance 34, 401–414. (1979) Harrison, J. M. & Kreps, D. M., Martingales and arbitrage in multiperiod securities markets, J. Economic Theory 20, 381–408. (1981) Harrison, J. M. & Pliska, S. R., Martingales and stochastic integrals in the theory of continuous trading, Stochastic Processes Applications 11, 215–260. (1983) Harrison, J. M. & Pliska, S. R., A stochastic calculus model of continuous trading: complete markets, Stochastic Processes Applications 15, 313–316. (1990) Heath, D., Jarrow, R. & Morton, A., Bond pricing and the term structure of interest rates: a discrete time approximation, J. Financial Quantitative Analysis 25, 419–440. (1992) Heath, D., Jarrow, R. & Morton, A., Bond pricing and the term structure of interest rates: a

CHAPTER 35. Notes and References
new methodology, Econometrica 60, 77–105.

359

(1993) Hogan, M., Problems in certain two-factor term structure models, Annals Applied Probability 3, 576–581. (1986) Ho, T. S. Y. & Lee, S.-B., Term structure movements and pricing interest rate contingent claims, J. Finance 41, 1011–1029. (1988) Huang, C.-f. & Litzenberger, R., Foundations for Financial Economics, North Holland, Amsterdam. (1993) Hull, J., Options, Futures, and other Derivative Securities, Prentice Hall, Englewood Cliffs, NJ. (1990) Hull, J. & White, A., Pricing interest rate derivative securities, Rev. Financial Studies 3, 573–592. (1994a) Hull, J. & White, A., Numerical procedures for implementing term structure models I: single-factor models, J. Derivatives, Fall, 7–16. (1994b) Hull, J. & White, A., Numerical procedures for implementing term structure models II: two factor models, J. Derviatives, Winter. (1987) Ingersoll, J. E., Theory of Financial Decision Making, Rowman & Littlefield. (1944) Itˆ , K., Stochastic integral, Proc. Imperial Acad. Tokyo 20, 519–524. o (1992) Jacka, S., A martingale representation result and an application to incomplete financial markets, Math. Finance 2, 239–250. (1990) Jaillet, P., Lamberton, D. & Lapeyre, B., Variational inequaltiies and the pricing of American options, Acta Appl. Math. 21, 263–289. (1989) Jamshidian, F., An exact bond option pricing formula, J. Finance 44, 205–209. (1990) Jamshidian, F., The preference-free determination of bond and option prices from the spot interest rate, Advances Futures & Options Research 4, 51–67. (1993a) Jamshidian, F., Price differentials, Risk Magazine 6/7.

360 (1993b) Jamshidian, F., Corralling quantos, Risk Magazine 7/3. (1996) Jamshidian, F., Sorting out swaptions, Risk Magazine 9/3, 59–60. (1988) Jarrow, R., Finance Theory, Prentice Hall, Englewood Cliffs, NJ. (1991a) Jarrow, R. & Madan, D., A characterization of complete security markets on a Brownian filtration, Math. Finance 1(3), 31–43. (1991b) Jarrow, R. & Madan, D., Valuing and hedging contingent claims on semimartingales, preprint, Johnson School of Business, Cornell University. (1991c) Jarrow, R. & Madan, D., Option pricing using the term structure of interest rates to hedge systematic discontinuities in asset returns, preprint, Johnson School of Business, Cornell University. (1981) Jarrow, R. & Oldfield, G., Forward contracts and futures contracts, J. Financial Economics 9, 373–382. (1983) Johnson, H., An analytic approximation for the American put price, J. Financial Quant. Analysis 18, 141–148. (1984) Jones, E. P., Option arbitrage and strategy with large price changes, J. Financial Economics 13, 91–113. (1951) Kac, M., On some connections between probability theory and differential and integral equations, Proc. 2nd Berkeley Symposium on Math. Stat. & Probability, 189–215. (1988) Karatzas, I., On the pricing of American options, App. Math. Optimization 17, 37–60. (1991) Karatzas, I. & Shreve, S. E., Brownian Motion and Stochastic Calculus, Springer-Verlag, New York. (1981) Karlin, S. & Taylor, H. M., A Second Course in Stochastic Processes, Academic Press, Orlando, FL. (1953) Kendall, M. G., Analysis of economic time-series – Part I: prices, J. Royal Statistical Society 96, Part I, 11–25. Reprinted in Cootner (1964), 85–99. (1992) Kloeden, P. E. & Platen, E., Numerical Solution of Stochastic Differential Equations, SpringerVerlag, New York.

A. D. Chelsea. M. 7. N. (1993) Lamberton. (1965) McKean. D. H. F. Option pricing with V. Ergebnisse Mathematik 2. (1986) MacMillan.CHAPTER 35.. P. (1993) Lakner. e e (1948) L´ vy.. New York. Analytic approximation for the American put option. S. e (1992a) Longstaff. (1976) Margrabe. S.. 16–23. Math. (1992) Kunitomo... Convergence of the critical price in the approximation of American options. Economic Literature 27. University of Pennsylvania.. J. G. On square-integrable martingales. 119–139.. A. Fixed Income 3. & Watanabe. & Schwartz. A theory of forward and futures prices. Grundbegriffe der Wahrscheinlichkeitsrechnung. J. F. Adv. The V. J. Finance 3. G.. E. P. Industrial Management Review . Efficient capital markets and martingales. & Ikeda. (1939) L´ vy. 511–524. Journal 30. 1259–1282. Paris. S. Sur certains processus stochastiques homog´ nes. (1992b) Longstaff. F. N. 1583– 1621. Math. Futures Options Research 1. Wharton School of Business. & Seneta. H. Compositio Math. & Schwartz. 43–53. Finance 47.. E. Pricing options with curved boundaries. F... 283–339... Gauthier-Villars. (1989) LeRoy. Nagoya Math. Martingale measures for a class of right-continuous processes. (1967) Kunita. 1950. Mathematical Finance 2. (1990) Madan. A. Notes and References 361 (1933) Kolmogorov. Interest rate volatility and the term structure: a twofactor general equilibrium model. P. J. Finance 3. W. English translation: Foundations of Probability Theory.. preprint.. A two-factor interest rate model and contingent claim valuation. A. Finance 1(4). L. Business 63. 275–298. Processus Stochastiques et Mouvement Brownian. E. (1991) Madan. 179–190. P. martingale components. Jr. 39–56. D. Appendix to P. S. & Milne. Samuelson (1965): A free boundary problem for the heat equation arising from a problem in mathematical economics. 209–245.. Math. model for share returns.

Operations Research 7. P.. 1–23. Arbitrage Pricing of Derivative Securities. Bell J. Rev. Futures Option Research 2. M.. M. (1987) Omberg.362 6. The pricing of the American option. Management Science 4. General equilibrium pricing of options on the market portfolio with discontinuous returns. 191–200.. New York. M. & Rutkowski. R. u Economic Letters 31 37–41. (1995) Øksendal. M. & Lee. (1990) Merton. & Runggaldier. R. 1990. 125–144. J. If the cap fits : : : . Lifetime portfolio selection under uncertainty: the continuous-time case. (1989) M¨ ller. (1990) Naik. 247–257. B. J. Option pricing when underlying stock returns are discontinuous. C. N. (1993) Mercurio. Cambridge. Econ. Ann. (1969) Merton. (1995) Reed. Option pricing: the American put. Option pricing for jump diffusions: approximations and their interpretation.. Stochastic Differential Equations. C.. The valuation of American puts with exponential exercise policies. On complete securities markets and the martingale property on securities prices. R.. Financial Studies 3.. Reprinted in Cootner (1964). Springer-Verlag. M. Brownian motion in the stock market. R. 4th Edition. Fin. 21–36. (1990) Protter. Review Economics Statistics 51... (1996) Musiela. W. 145–173. Stochastic Integration and Differential Equations. Economics 3. Adv. 34–35. F. Risk 8(8). (1959) Osborne. New York. to appear. Springer-Verlag. Business 50. (1992) Myneni.. (1976) Merton. (1977) Parkinson. Finance 3. R. Applied Probability 2.. Theory and Applications.. Theory of rational option pricing. New York. C. Springer-Verlag. (1973) Merton. 117–142.. F. M. Continuous-Time Finance. 32–39. Basil Blackwell. 100–128. V. Math. 493–521. C. E. S.. 141– 183. .

1–25. Risk Magazine 4/8.. & Shi. New York. (1961) Sprenkle. Warrant prices as indicators of expectations and preferences. Inst. M. M. 177–188. Yale Economic Essays 1. P. (1991) Shirakawa.. & Reiner. M. (1987) Ritchken. W. Duffie. (1965) Samuelson. P... The value of an Asian option. (1995) Rogers.. J. Strategy and Applications.. C. J.CHAPTER 35. L. M. & Willinger. O. Martingale densities for general asset prices. (M. Appl. Breaking down the barriers. (1992b) Schweizer. P. 28–35.. Financial Economics 5. Ann.. M.. Financial Economics. An equilibrium characterization of the term structure. Math. 93–116. 171– 179. Finance 1(4). Options: Theory.. A. Reprinted in Cootner (1964). e (1987) Taqqu. (1990) Stricker. E. IMA Vol. C. G. & Sundaresan. 178–231. G. Interest rate option pricing with Poisson-Gaussian forward rate curve processes.. 123– 131. 412–474. 41–50. J. Adv. The analysis of finite security markets using martingales. C. F. J. to appear. Which model of term-structure of interest rates should one use?. 451–460. 77–94.. SIAM Review 15. 19. henri Poincar´ 26. Probab. Mean-variance hedging for general claims. (1991) Rubinstein. Springer-Verlag. Davis. W. H. (1992a) Schweizer. .. 65: Mathematical Finance. Probab. Annals Appl. A. (1973) Samuelson. 2. (1995) Rogers. Industrial Management Review 4. Mathematics of speculative prices. L. editors). Economics 21.. 1–42. Optimal dividend and portfolio decisions with Poisson and diffusion-type return processes. Proof that properly anticipated prices fluctuate randomly. Math. D. Harper Collins. Z. Notes and References 363 (1981) Richard. Tokyo Institute of Technology. H. 9. preprint IHSS 90-20. Shreve. Arbitrage et lois de martingales. Applied Probability. 347–371. (1990) Shirakawa. C. S. M. (1977) Vasicek. Fleming & S. A continuous-time equilibrium model of forward and futures prices in a multigood economy.

(1923) Wiener. France 52. N. (1995) Wilmott. UK. J. . The Mathematics of Financial Derivatives. J. (1991) Williams. and its applications. Un probl´ me de probabilit´ s d´ nombrables. Math. Differential spaces. Cambridge. A survey of stochastic continuous time models of the term structure of interest rates. D. X. Paris. Cambridge University Press. Insurance: Mathematics and Economics 14.. P. N. Martingale representation for a class of processes with independent increments.. Etude Critique de la Notion du Collectif. Berlin. Bull. X. (1924) Wiener. S. Soc. OX4 1DR... Dewynne. R.. Dewynne. K.. Probability with Martingales. Oxford. & Howison. 569– e e e 578. Physics 2. ´ (1939) Ville. SpringerVerlag. & Howison. J. Math. Option Pricing. 131–174... 279–311. (1992) Xue. (1993) Wilmott. Gauthier-Villars. S... Cambridge University Press. 139–161.364 (1994) Vetzal. J. in Lecture Notes in Control and Information Sciences 177. PO Box 348. Oxford Financial Press. Cambridge. P.

Sign up to vote on this title
UsefulNot useful