Lesson 23 Modeling Stock Prices with the Lognormal Distribution Reading: Derientves Markets 18.1-18.4 ‘Starting with this lesson, the normal distribution will be used for many of our calculations. See the| introduction, page x for & discussion about how you will be doing these calculations when you 2D take the exam, and how they will be dene in this manual 23.1 The normal and lognormal distributions 23.1.1. The normal distribution We are going to use the normal distsibution in our model for stock prices. Let's review the properties of the normal distribution, ‘The normal distribution is a two-parameter distribution. Usually, the letters pp and of are used for the parameters. However, we will be using s for a different (although related) purpose, so we shall use the letters rand v as the paremeters instead. The formula sheet you gett the exam will use mando as wel ‘The probability density function of a normal distribution with parameters m: and 0 is Fain) = Itis not necessary to memorize his function since its in the tables given to you at the exam. ‘The mean is and the variance is v2, The ranciom variable having this distribution is said to have an Nm, 8?) distribution. Notice that the second parameter is squared. "A sedend normal distrivation has» ~ 0,2. ~ 1. Thepprokability density function at z fora standard normal distribution is denoted by 412) ‘To evaluate the cumulauve distribution function, we tanslate x into the cosresponding value fora standard normal distribution a follows where N(] is thestandard normal distribution function at =. We then ean utilize normal distribution function ‘provided on some calculators, or a function provided in many spreadsheots such as NORMSDIST in Excel, to {aleulate the value of the standard normal distribution function." Exanrte 23. X follows an A(50, 36) distribation. Caleslate Pr{X < 59), Fein.) Se prides NORVIDIST clo asay Mana ton ping su Copp eas ASL322 23. MODELING STOCK PRICES WITH THE LOGNORMAL DISTRIBUTION Pr(X 5 58) = F(G9) = ne) =N«5) VB o Here sabre review of other properties the normal stbution Ifyou need more deta elther refer to the MeDonald tebook orto yout favorite probability textbook + The normal distrbution is symmetric around its mean. ‘+ The normal distribution isa sable distribution, This means that ifa set of normal random vastebles forms 2 multivariate normal distribution, any linear combination of those variables ie normal. In particular, the sum of independent normal variables is normal '* The limit of the sum of independent identically distributed random variables, regardless of their disti- ution, is normal. This isa version of the Centra! Limit Theerem. 23.1.2. The lognormal distribution Let X be @ normal random yatiable with parameters m and , and let Y = e®. Then Y has a lognormal distribution with parameters m and v. Conversely if Y has 2 lognormal distribution with parameters m and , then X = In Y has's normal dstribation with parameters mand >. The probability density function fora lognormal with parameters mr and 9 18 ete ntitet Sime) A lognormal distribution is skewed to the right. While the median ofa lognormal ise”, the made is ‘The proof of tis sin Table 23.1.2 “The mean of a lognormal madom veriable X with parameters wand v is px] = ease ey ‘You are responsible for knowing this fact, ut not its proof. If you are curious about the proof, you ean find it In the textbook’s Appendix 18.4, Alternatively, the derivation of the partial expectation of a lognormal is given In Table 232 on page 331) set k = o> in that decivation to obtain the derivation of the mean. The texibook mentiors that the variance of the lognormal random variable X is -1) ‘The page before the tables that you get at the exam provides the following formula for the kth moment of a lognormal: va =" EX") Multiplying lognormal random variables ith parameters 1m and 0 i simlae to adding normal random variables. I the lognormal random Variables ar independent, the product isa lognormal random variable having parameters mt = [mand o* = 5.02, Wal ofthe multiplied variabies have the seme parameters, the product of n of them has m = min and 9 = 6:VA. “Te MeDorald tottook Goes ot ste the wads of lognormal, so you are not respenibe fr never it Howe, NeDonul mentors the chapter summary (ope 619) ht thelegneeral Sateen i “urimedal (Fa ene hump), which eit eet 0 pg,231. THE NORMAL AND LOGNORMAL DISTRIBUTIONS 333 ‘Table 23. he formula forthe made of ¢ lognormal ‘The probability density function of a lognormal is pctuscnP at ae flem.2) io determine the maximum with respect 10. we can drop the mulupeatwe constant 1 (ov the function, since logging doos not change tho maximum. So itsarfices maximize |. We can log gia) Difcrentinte and set equal to 0. x= Multiply through by 0%, sins im =o Inx=m—o ‘Thue the mode of a lognormal random variable se" 23.1.3 Jensen's inequality We have just observed that the oxpected value of an exponentiated normal random variable is greater than the exponential ofthe expected value of that normal random variable, ‘More generally, convex function g(x) is defined as one having the following property 1 g(ex+(1-a)y) sags) +(-algty) for 0s In other words, fr alls, the value of g(x) is less than the inear interpolation of ¢(x:)and gi22) for z= x $ 32. ‘Any function with a positive second destvative everywhere, such as g(x) = e¥, a convex function. Ther: ‘Theorem 1 Jensen's inequality) IfX iso rnd variable and g(x) sa conmex fnctio, then F[s9) = s(818]) ° For example, consider a discrete random variable X with the following distribution: 103 bts Then the expected value of X i$ 0.5(1 + 5) = 3. Consider the function g(z) = 2%, or squaring the random varlable. The expected value of X? is 05(1 +25) = 13, but the square ofthe expected value of X is 9. We see that 13 >9. "The proof of Jonson's inequality goes as fllows: nsSuty Nan" don 2 pring Cpaatemie aa3 23. MODELING STOCK PRICES WITH THE LOGNORMAL DISTRIBUTION Poor Since s(x) convex, possible to draw a line tangent to its graph at any polnt x. Ths tangent ine willbe less than or equal t the graph atall points. Draw sucha line at the point x = E[X]. Call the line L(). ‘Then: Statement Reason E[z00] > E[Leo] | gt) > Lia) atall paints x by convexity F[L¢ei] = L(ELXI) Expectation is linear L(EIX) = g(81X1) | Thelline L(x) is tangent to (+) atx = EIX] Putting the threo statements together yields Jonsen’s inoquality . 23.2 The lognormal distribution as a model for stock prices 23.2.1 Stocks without dividends ‘Weill temporarity limit the discussion to stocks that pay no dividends, Let 5, be the price of a stock at time . in our model of stock prices, we are going to assume that the rate of return on the stock is normally distributed. Suppose a isthe continuously compounded annual rate of return. If there were no randomness, then we would have §; = Sie, or 54/59 ~ e#. Since $, is random, we instead hhave 5,59 = e4, whore A ie a normally distributed random variable, Then S,/S) is lognormally distributed and E{5,/S0] = Ele*] = e*!. The mean of $;/S0 is the mean of the lognormal random variable. ‘The definition of the rate of return can be confusing. Let’s consider a simpler (non-normal) example. Suppose an investment of 100 at time 0 will result, after @ year, in 50 with probability 0.5 and 200 with protability 05, What is the rate of return on this investment? There are two ways of looking at it 1, The average of 200 and 50, or the expected valve of the investment after one yest; Is 125, Therefore the sfiective annul rate of retum is 125/100— 1 = 25%, and the continuously compounded annual rate of return is In 1.2. 2, Ifyou get 200, the rate of return Is the number a such that 1008 = 200, so a = In2. Ifyou get $0, the rate of return is the number 22 such that 1002 = 50, so a2 = ~In2. The average of a; are az is 0, 0 the rate of return is 0%, Which way is correct? Almost anyone would consider the fest way correct. Because of Jensen's inequality, the expected valuas of 1 and e are higher than the exponentials ofthe expected valos of 4 and a Therefore, when we refer to the continucusly compounded annual rate of return 2, we ae referring tothe logarithm of the mean of the lognormal random vanable $:/So: a = In E[S:/S)]. If this lognormal random variable has parameters m and v, then by equation (231), a = m+0.502, we need to refer to the mean of In(Si/S;), or m, we wil eal tthe expected value ofthe annual rate of retuan, in contradistinction {0 the ontindously compounded anna rate of return which is Let's generalize from a 1-year period to aI-year period. In our lognormal model if the parameters af the lognormal distnbution for 53/50 are m = j= @~0.50" and o = o, then the parameters of the lognormal random variable $;/S) are mt = ur = at ~ 05s and 2 = oF. Generally, you will be given the following parameters ofthe lognormal tock model ‘+ c; the continuously compounded annual rate of return + othe (annual) volatility [stay Mina atin pring lpg eons a23.2. THE LOGNOKMAL DISTRIBUTION AS A MODEL FOR STOCK PRICES 25 Bxanerue 298 The time+ price ofa stock is). You are given (i) The continuously compounded annual rate of retum on the stock is 0.15. (i) The stock's volatility s 0.3. (Ui) The stock pays no dividends. lav) 5 = 80 Caleulate the probability that Sis at least 150. “Axewnr: Si/So 62 lognormal random variable with parameters mi = (« —0.50")! = (0.15 -0.5(037))(4) = 042 and 9 = oV! = (34 = 06. Therefore l(Sx/So) is normal with the same parameters. We want PriSa/So > 150/80): 15 5] ‘To slustiate the lognormal model for stock prices, suppose the parameters of the stock price mociel are: -=045and o = 03,90 that In; So) is normal with parameters ot = 0.15 and @ = 03. Figure 231 graphs the probability density function of (eon 0, 2) 06 a {@) the normal random variable with parameters m andl 2, and {b) the stock pric ser one year, ‘You will notice that while thenormal random variables symmetric arcund 0.15, the stock prise graph isnot symmetric around 40e™ = 46.47, Exponentictinga random variable preserves its percentiles, so the median stock price is 45,47. However, this snot the most likely stock price; 40e°5-0°" = 42.47 is the most likely stock pce, Moreover the expected sock pices 4 (e020) = a6), whichis greater tan the median. Thee fs grater thar 50% chance Bataleon random oavoble el be less han ts mea. ExaMPLt 23C 5) Is the price of a nondividend paying stock at time tS; follows a lognormal model, You are iver: (0 S0=40. () Thestock’s continuously compoundes expected growth rates « = 0.15; (ii) Thestock’s volatility » = 03. Determine 1. The average price ofthe stork after one year 2. The average price o the stock after four yea. 3. Tho medion pric o the stock after one year, 4. The median price of th 5. E[ln(Ss/5s)] 6. Eltn(ss/S0)} ‘after four years. foam: ite goonngci ot ile compounded expected growth rate, [Si] TE ocige tet ote ek cae Yous ont OR oe =i, juestions. By definition of continuously Ta Say Manual tan pting Cant emb Aa326 23 MODELING STOCK PRICES WITH THE LOGNORMAL DISTRIBUTION 9(#10.15,0.3) o7/ 06 os oa os 02] oa as 06 —0a_—33 ~~ SSCS (@)Potabity drt function forthe cervnuously compounded axl vate of return (S/S) Probability Density 0.03 0.025 02 os] oot (0.005 a a a a a a Probl dens fncon othe ck price afer one eS Figure 23.1: Graphs ofthe rate o!relurn and the stock price aiter one year under a lognormal model. The parameters ) w= 15, 0 = 03. 3, The lognormal parameter is m = p= a 0502 = 0.15 ~ after one years Se = 0.8" - [agane4) 4. The lognormal parameter ism = 41 ~ 4(0.105) = 0.4, and the median price ofthe stock after four years is Se = cast? Since v = oVl = 0.6, the average price ofthe stock after four years can also be caleulatod as Soo" = Sot21050.) 49.05, which isthe samo as the solution above. 5. Logging the quotient recovers the normal random variable with mean js = a ~ 50? and variance a2, so the average I-year growth rate is = 0.15 ~D 510°) = 6. The average d-yosr growth ratis Aj = 40.105) ~ a .5(03") ~ 0.105. The median price of the tock 23.2.2 Stocks with dividends ‘To incorporate dividends into a continuous lognormal model, we assume that the stock pays dividends at a continuous rate proportional tots price. Denote the annual rate by 5. The total rate of return of the stock isc, bbut part of this return ie paid out asa dividend. The price ofthe stock increases atthe rate a — 8) pase Mansa esa" paar,232, THE LOGNORMAL DISTRIBUTION AS A MODEL FOR STOCK PRICES 327 «The continuously compounded rae of stock price appreciation isa ~ 6. The ratio S/S) is @ lognormal random variable with parameters m = (230.532)! andv = 0 Vi. Thenetrate afsiock priceappreciation fr isalee known a8 the “capital gaits rate”. «The continuously compounded annual rate of return on the stock is a. 1X; represents the cumulative Aalucer an inesnent of Spin the stock with dividends confinwonsly reinvested, then X;/Xo 1s lognormal with parameters m = (@ -0.502)! and 0 = oF. Exawets 25D The time+t price ofa stock is Sy. You are given () The continuously compounded annual rate of return an the stock is 0.15. (@) The stock's wolatity is 0.3. (ii) The stock pays dividends proportional tots price et an annual rate of 002 liv) Se = 80 Caleulate the probability that Si is atleast 150. [Answers ‘The parameters of the lognormal random variable S4/So are 1m =(015-0.0-051037))4) 03 =06 ost “The dosied probabiity ts a n(& ‘Quiz 25-4 A stock's price 5; follows a lognormal model. You are given! @) $= 100. ii) The continuously compounded expected rate of return is 0.12, (i) The continuously compounded dividend rote is 0.3. Gy) The volanlity is. Determine the value # for which the median price ofthe stock a time tis 110. In(450/50)- 0.94) _ 1 — (0.48101) = [031525] 23.2.3. Prediction intervals Using the lognormal model, we can answer questions on the distnbutton of the prize ofthe stock, We can caleulate the probability thatthe price ofa stock isin ¢ certain range. We ean also calculate prediction intervels* for the price of the stock. The term “p-prediction interval” connotes an interral centered at the mean of random variable for which p is the probability thatthe random, vvaricble is inthe interval ExAMrie 256 A stock’s prices follow o lognormal distribution, You are given: wa wo au) o 1. Detetmine the probability thatthe stock’s price atthe end of one month willbe greater than its current price Gate score Sion of Doeatin Wavkls, McDoeald mistaken ells that intra “emfidenee neva" 0 yea may fn he term anfdenenteerde” sedan ad ens one eanib A elon pet38 23, MODELING STOCK PRICES WITH THE LOGNORMAL DISTRIBUTION 2. Determine the probability that the stock’s price at the end of one month willbe greater than the expected price 3. Construct « 95% symmetric prediction interval for the ratio of the stock's price at the end of a manth to the current price. Answer: 1. For one month, the parameters of the lognormal distribution are 0.49.02 - 050.) aT 00625 Hla 5-050°) and » = 03/1712. The probability thet 5.2 > S oF Sia(S0 > 1s the probability that lS, 2/59) > 0, oF that the corresponding normal distribution is grester than. This is vy {@=a.00625| o3yi7ia (0.01, the mean 2 Now we want the probability thatthe growth in price willbe greater than 4 =) growth in price (0.0 ae) ed ( oayin 3, A 95% prediction interval is an interval centered at the mean and having probability 95%. ‘The lower bbound of the interval i the 25 percentile of the distrbution and the upper bound 1s the 97.5 percentile, For a standand normal distribution, the 975 percentile is 1.96. Therefore, we add and subtract 1.96 times = 6VE = 03yf712 to m, and thon oxponentiaethe result. 0317 = 003660 0.00625 + 1.96(0.086603) = 0.17599 e179 = 1.19245 (0.00625 ~ 1.96(0.086606) = -0. 16349 hee 0.81917 ‘The answer () The stock's current price is 80 (i) The stock’s expected continously compounded rate of rebum ie 0.1. (ii) The stock's volaiity is 0.25 Constnact a 95% prediction interval forthe stock's prico aftor one year using McDonald's method. @) (siz 25-2 A nondividend paying sack follows a lognormal model You are given 23.3 Conditional payoffs using the lognormal model ‘The standard norinal cumulative distribution function at x Is the probability that a standard normal random variable X is les than or equal to x. In other contexts, it is denoted by 2), but in financial economics it is traditional to denote ity N(x). You are given a table of N(x) values at the exam {BuLSay Yanan tin 2 penne Coeur ea233. CONDITIONAL PAYOFFS USING THE LOGNORMAL MODEL. a9 ‘The payoff on a European opticn is conditional en the stock price being above or below a ceriain price at a certain time. In this section, we will calculate expected payoifs of options, For the remainder of this section, ail options are Enropean options. “Asa first sep, Wve cen calculate the probability that an option will pay off. Suppose the price ofa stock is Syyard a, 6,and.v have the usual meanings. Let the strike price of a put option expiring at fime f be K. Then the put option pays off §, < K. The lognormal distribution of S/S has parameters m = (a ~6—0.50°)t and avi, s0 axe eo where /K)+(w-8-05e%t ove 5; > K, and the probability of that is the complement of the above Now consider call option. Itpays off prababiity whichis 1— Ni) Pr(S; > K) = N(da) (33) ‘Quiz 23-3. A stock's price followsa lognormal model. You are given - @) = 00 i) a= 045 (uy 0 = 02 Gv) 8 = 0.05 ‘A European call option on the stock with stitke price 70 expires in 3 months. Calculate the probability that the option pays off Next we want to calculate the conditional payoff of the option, given that it pays olf. Let's start with the following definition: Dininrri0N 23.A The partial expectation of a continuous random variable X having probability density func fion j(2), given that i isin the interval [a,b] (either bound of the interval may be infinite), Is the eontsibution. to the expectation from values in the interval [a,b], or [fx [A suiniler definition may be made for discrete random variables. In other words, the partial expectation is the piece ofthe total expeciation over an interval. The sum of the partial expectations of a random variable over all intervals from ~co to eis the variebie's expected valve, For example, considera fair sivsided die, The partial expectation of a toss given that the toss is less than 4 is (1+2+3)/6 = 1. The partial expectation of a toss given that the toss is greater than or equal to 4 is [4415 - 6)/6 ~2.5. The sum of the two partial expectations is the expectation of a toss, or 3.5. This manual will use the notation PELX | Y] for the partial expectation of X given ¥. This notation is not used in McDonald and will not appear on the exer, ut will be convenient ‘By dividing the partial expectation conditional on an interval by the probability ofthe interval, you obtain the conditional expectation conditional on the interval, In other words: PEIX | ¥) Pry) Bx |Y1= (Bay Trusty Mt atm" png330 23. MODELING STOCK PRICES WITH THE LOGNORMAL DISTRIBUTION ‘The partial expectation of a lognormal random variable with parameters m and given that itis less then i pepe x enya E==) ‘sfondi e232. MaDonald ele eon of hs frase at yeu Pickin sia wk aun nomen coerniae {Opp tomes see tha 5) cage Te probably Py < K) the somes ‘Pr(S;/ Sy < K/ Sp), 80 let k = K/Sq in the formula. The partial expectation of 5; ts Sp times the partial expectation ‘of 5;/Sp. We then have 5) PEIS:| 5; < I= eens (BRIS) = Se (Ao ln Fae) = Soe (di) 35) where InfSe/K) + (e-— 8+ 050%) ont ‘From equation (23.2), Pr(S; < K) = N(-dp). It follows from equations (23.5) and (23.4) that See“9# Nid) EIS) 15, < kj = BON) 23.5) S18 Kis the full expectation of &; minus the partial expectation ofS; given that 5, < K, which we calculated above in formula (23.5). We get ES | 5) > K]~ Soe! — Se! PHN) = Syl) es) oe NG ‘One who owns a European call eption will ave the following cash flows at expiry EIS) | 5: > KI (34) {Fusing Maa etn pg Copongn cas a283, CONDITIONAL PAYOFFS USING THE LOGNORMAL MODEL, a “Table 2.2: Derivation of Partial Expectation of Lognermal SW wat pana SAT ronal clon Vaal Kvh pate and ven al ee Rae bhi deny fener ctalogromal wih parameters mando meetne La aa a “Therefore, the paral expectation i PELX |X <&] be ge eal Anse vie Substitute y = Inx—m: zoe ax =eh™ay |As part of the substitution, translate the bounds ofthe integral from x to nx — o> ko Ink—m and the expression becomes 1 ee pay, oe fo Pees EP Paernay = erates Win d= 4 iE Jo S =P te teetay wad. ‘i We wart the numerator of the contined exponent to bo a negative square To complete the square, we need to turn y! —20%y into (y ~ B) for some b. Wehave 2by = 20%y, sob = oF and yo doty ‘The resulting expression forthe partial expectation becomes yo e ort Raya enese [" A_ et atay Win dow vin |The integrand is now the density of a normal distribution with mean o* and variance 0 so the mtegra the probability shat such a variable less than In e— m1, which i N (nk ~m ~ 0°)/2). The partial expeciation is therofoce k-m=- Ea paix [x Kis Soet9NCA) wid) ‘The probability of this receipt is Nida). Hence, the expected value of this receipt is Soe! N(ds) HIS: 15) > K Therefore, the expected payoff on a call option is Efman(0, 5; — Ky] = Syel*-'N()— KNIda) (23.10) ‘The formulas are summarized in Table 233. Feaurre23F A stock's price follows s lognormal distribution. You are given ‘The stock's price is currently 30. Determine the conditional expected value of the stock's price after 3 months, given that itishigher than 75. Answer: We will use formulas 233) and (23.8). First lets calculate dy and ds, Notice that in general, th = dy oe. (60/75) + 0.15 +0.3*/2)(0-25) 03vo25 =2.37810 and the answer is PE(S025 | Sax5 > 751 Pr(So2s > 75) \N(=237810) a spl 04029) me ‘N@252810) eon) susie BSz5 | Saas > 75] (Quiz 23-4 Fora nondividend paying stock, youare given (i) The stock price follows a lognormal model (6) The curtent price is 100. (bi). The continuously compounded expected rate of return is 01 (iv) The volatility is 0.2. A European call option on tho stock expiring in one year has strike price 100, Calculate the expected payott on the cal option. To calculate the value of an eption, we need to discount the expected payoif of the option. But that isa problem. What is the appropriate discount rate? This problem is solved the same way we solved the problers TEMS Mort ena pnts Cope Sa ao23.3. CONDITIONAL PAYOFFS USING THE LOGNORMAL MODEL 333 Table 23.8: Formula Summary for Lesson 20 For a slock whose price 5; follows a lognormal model: + The expected value is EIS | So] = Soe" @3.)| + di and dare defined by y= HASoIKD+ ta 5 + 0502) ow tae d—ovt + Probabilities of payoffs and partial expectations of stock prices are Pr(S: < K)= NA) (3.2) Pas; > K) = NG) (33) ELS; | 5: < K] = Sue*NC-e) 35) Soet*-SIN=dd BIS) |S) < K]= 2 236) Isicxl= "Sy (236) EIS: | 81 > K] = Soe! #N(A) (3s) soot OONCE) Eis, 1, > x)= SN 239) Isox= oF c + Expected option payotts are = Cal: ; ; E{maxi0, 5; - K)] = Soe™'N(d) — KN (A 23.10) = Put Elmax(0, K—S)]] = KN) - See NA) eu) {for binomial trees. Rather than using the true lognormal distribution to project stock values, we tse the risk neutral distcibutien, wich uses rin place of , ond discount the riskeneutral expected value using the risk-free rate. Doing this gives the Black-Scholes formula, which we will learn in 2 later lesson. The BlackScholes formula uses unhatted dy and da, which are defined with r's instead of a's ‘So rather than memorizing formulas (23.10) anel 23.7), you should memorize the Black-Scholes formula, and know hw to go from the Black-Scholes formula to these formulas: accumulate to the end of the period fanc then replace with a. AS Meat itn peg Capon canis aewe 23, MODELING STOCK PRICES WITH THE LOGNORMAL Di Exercises ‘Stock prices n the lognormal model 23,1. Foranondividdend paying stock: (‘The price of the stcck at ime fis S, (i) The annual continuously compourcled expected rate of return ais 10%. (i) The median of the stock price at time 6 Spe! Determine the volatility of the stock. 232. A stock's price follows. lognormal model. You are given: (‘The annual continuously compounded expected return on the stock i 10%, (i) The volatility ofthe stock i 30%, Determine the probability that the rate of return over a one year period willbe less than the expected rate of return, 23.3. A stock's price follows a lognormal model. You are given’ (The current price ofthe stock is 100. (ii) The probability that the stock’s price will be less than 95 at the end of 3 months is 0.2858, (Gi) ‘The probability that the stock's price willbe less than 110 at the end of 6 months is 0.6025. Calculate the expected price ofthe stock at the end of one yeer 2B4. A stock’s price followsa lognormal model. You are given’ (The initiel price is 100 i) a=0a5 (ai) 5 = 0.05 (iw) o= 05 Construct a 95% prediction interval forthe price ofthe stock atthe ond af five years 23.5. The price ofa stock follows a lognormal medel. The current price ofthe steck i 4. ‘A95% prediction interval for the stock price at the end of one year is (40.10, 62.05). Construct a 90% prediction interval forthe stock price at the end of two years. 23.5, Astoek’s price follows a lognormal model. You are given: w=0a5 (uy 5 =0.02 (i) 3 Determine the smallest period # > O ouch that the probability that the stock price at tis less than the initial price is less than 5%. ny Mn ain pining Entec om ft Copyngh cab ke easEXERCISES FOR LESSON 25 335 se thejlleeing information jor questions 25,7 aad 23.8: Wri the price of a nondividend paying stock at time f. §; follows a lognormal model, You are givens @ So=45. Gi) The continuously compounded expected rate of retusn 1 0.15. (ii) The volatility is 02 237. "Theprobability thatthe stock price Is less than 40 at time fis 0.10. Determine all possible values for 238, Let p; be the probability thet ; is less than 40. Determine the value off for which pis maximized overall postive! 239, [CAS8-50030] You are given the following information about a European call option. ‘The current stack price is 35. "The oxereise price is 40. “The option matures in 6 months. ‘The expected annual return on the stock is 18% ‘The annwal volatility ofthe stock price is 24%. ‘The stock's price at each future time T, given its current price, is lgnormally distributed. ‘The stock pays no dividends. Determine the probability thatthe call option will be exercised. 2210. A steck's price follows a lognormal distribution. To simulate its price over 10 years, scenarios are generated. In each scenario, the stock price al tne fis genevated by generating » standard normal random Yarlable Z. Then 5; is set equal to S,-re8!*02 for t= 1,2,.../10. Determane the expected value of the ratio 50/5) in this simulation. Conditional payoits 23,11. A stock’ price follows a lognormal model. You are givens (@) Thestock's initial price is 4. ()_ The stock's continuously compounded rate of return is 0.05. (it). Thesteck’s continuously compounded dividend raie is 0.05. Gv) Volauity 803, Calculate the conditional expected value of the stock after 1 year given that itis greater than 40. 23:12, A stock's price follows a lognormal model, You are givers (i) The stock's initial price i 40, (i) e020 (ii) 9 = 0.15 (iv) The stock pays no dividends, Calculate the conditional expected value of the stock after 2 years given that iti less than 59, ASwly Malt ston pane Execs cantinueon dc et page Copp ona336 23, MODELING STOCK PRICES WITH THE LOGNORMAL DISTRIBUTION 23.13. A stock's price follows a lognormal model. You are given: (The stocks intial price i 60. i) 2-015 (aay) 502. (vy) 5=003 Calculate the conditional expected value ofthe stock after 6 months given that it is greater than 50, 25.14. A stock's price follows a lognormal model. You are given: (i) The stock's earrent price is 38. (ii) The stock’s continuously compounded expected rate of return is 0.14 (ii) The stock's continuously compounded dividend rate is 004. (Go) Tho stock’s wobstlty ie 0.28 ‘A European call option on the stock expires in 3 months and has strike price 40, Caleslate the expected payoff of the call option 23.15. You have written a Furepean call option on a nondividiend paying stock with strike price 50. The option expires in one year. You are given: (i) The stock's price follows a lognormal model (ii) The stock's current price is 40. (ii) The stock's continuously compounded expected rate of return is 015. (iv) Thestock's volatility is 0.3. Calculate TVak for the written option at the 5% level 23.16. You have wnittena European put option ona nondividend paying stock with strike price 50. The option expires in one yeer. You are given: (i) The stock's price follows a lognormal model. (i), The stock's curront price 50. (ii) The stock's continuously compounded expected rate of return is 0.12. (Gv) The stock's volatility is 0.2. Calculate TVaR for the written option atthe 10% level. Additional released exam questions: C-S07:4, Advanced Derivatives Sample:50 (C-SO7:4 ancl Advanced Derivatives Sample:50 are the same, Solutions 1s¢ the fact that the median and mean of a lognormal random variable dilfer in the exponent by 08 tymediian condition ty mean condition [pu sea Mara tian print, Expyngh ean SeEXERCISE SOLUTIONS FOR LESSON 23, 337 232, The rate of return has a normal dietribution. The parameter m of the normal distribution satisfies mi +050! ~0.1, 20 m = 0.1 05/032) = 0055. Then the quantile of the expected return is (0.15) = 0.58952 ‘The probability tha he rte of return wil be ess than expected is [33963] 233, The probability statoments result in these two equations expressing the percentiles in terms of standard znormal coefficients of the 100pth percentile, or zp: 100e8250+0Soaa98 = 95 00284 VESouses = 110, ‘Thenormal percentiles are Zoz395 =-0.72and ze a5 = 0.26. Wollog the equations: 0125. 0.367 = 1n0.95 05y-+0.26V05e = In Lt and solve them. Double the first and subtract from the second to eliminate (02685 +0.72)0 = 101.4 -21n095 = 0197007 187397 OAS7O —o21e9 = 4(ln0.95 + 0.36(0.2189)) = 0.1101 ‘The expected value ofthe stock aftr one year is 10001048802 234, ‘The lognomal parameter forthe dsttbution ofthe sock price ale five yeas (a ~ § ~ 0502) = (0.1 0425)6) = -0.125. The lognormal parameter « is 055 = 1.1160. The prediction interval is -0.125 + 1,96(1.1160) ~ (-2.3163,2.0660). Exponentinting, we get (e°2312, e066) ~ (0,0986, 78950), Multiplying by the Inka stock price of 00, the answer i (0.96, 76959) 285. Welknow that 40.10 = 450!“18° and 62.05 = 450%, So the quotients 08 eo = SE 1.5078 = (in 5474)/5.92 = 0.1114 and the product is 2 ~ 401016205) eS Weneed 45¢!?! 088 and 45020" 65¥%e, ibeBe p1assiFart1) 521 NES « 45(3,2087)/1.20 a5,28°1605VE0 ~ 45(1.2287)1.2057) = 71.64 ‘The answer is[(42.67,71.64)| enya i pig338 23. MODELING STOCK PRICES WITH THE LOGNORMAL DISTRIBUTION 23.6. The wand o parameters are u = 0.15 ~ 0.02 ~ 0.3°/2 = 0.085 and .3, The fifth percentile of a normal Aserbuon —1.55. The fith percentile of In(5/5) 18 .085¢ ~1.685(0.3V7). Ne want this tbe greter than 0.0 0085" 1.685(0.3V8 > 0, Solving: (.085¢ > 0.4985 o.085Vi > 04955, Vi 2038 ea) 23.2. Dy formule 232), Pr(S, < 40) Ings) — (0.15 ~0. a0) We want to solve fort. Let x= vi. =in1azs Oar 0.13%? ~ 0.25631 + 0.11783 0.25681 + VODOEEEE a6 0.532, 1544) = 07293,1.203 238. By formula 232), in(45/40)~ (015 - 0510.2") Davi } We want to maximize this expression. ‘To make the calculus a litle easier let x = Vi; maximizing overall xis equivalent to maximizing overall. Maximizing the argument of N will result in maximizing the probability Let f(z) be thatargument. Pris: < 40) = | ie) fe- 1 0.322)(0.2) Set the numerator equal 0, and divide stby 02 ~0262 naz In 1.125 “os Coma cannaEXERCISE SOLUTIONS FOR LESSON 23 339 299, The lognormal distibuton of S 4/5 hes parameters b= la 0502/05) 2avTS ‘The call will be exercised if Ses > 40,0 the probability oft being exercised is ~ 065350 - [036657] 0.18 -05(0.2#)) (05) ~ 0.0756 Pr{Sos > 40) — (0.34136) i dps 8) 02405 2030, For this lognormal distribution, = 01 and o = 02, om = jt = Jandy = oV = O2ViT. Themeanis 2BAL Wo use formula (239), jy = In(40/a0) + 0.05 — 01.05 + 0.503%) _ ce 2035 d,=d, avi =0.15-0.3 =-0.15 ES; | S) > 40] = Soe" Nd) ap NOS) = 8X0) oss3e2 = «(fans - I 2312, Weuse formala 226. 4 LAOISO 1010-01879) gga 0.15v2 dy = dy 0.15N2 = -0.21516 ona) Neds 2 ( 50] = See!) Uencsaty Mada etn 2 pnt Copragh esis a340 23. MODELING STOCK PRICES WITH THE LOGNORMAL DISTRIBUTION soc. ons 23.18 We se formula 23.10) j= INSBHO) + [04-004 + 051028025) O76 0.11781 0.28025 ~ 025781 Odi) = N(-0.11781) = 045811 N(d) = N(-0.25781) = 0.39603 The expected payotf of the cll option is Se NEG) - KNOG 0.41781 8° (0.45311) ~ 40(0.99828) (73) pete reheat to te aretha od eet ice rises the most, The normal paraareters are m = 015 ~0.5(0.37) = 0.105and v = 0:3. The 95" percentile Stthe tock price at the end of cre yearia 40 exp (0105 + 1.645(0.3)) = 72.7755 The expected price ofthe stock, given that ts price is at least 72.7755, is piven by equation (23.8), and Is j= (40/72.7755) + 0.15 + 0.5(0.8) 03 d= 145-03 = 1.645 N(d)= 0.08931 NGd:) = 0.08998 0029(0.08951) To Thus the expected an given thatthe stock price Is greater than 72.7755, i §3,040~50 = 93.0440. The TVR of the payof is[=33.0040] 28.16. Since you've walten the option, the worst case fs when the option pays the mest oF when the slock price falls the most. The normal parsineters are r= 0.12 ~ 0.5(02) = 0.10 and v = 0.2. The 10" percentile of the stock's price atthe end of one year is ‘S0exp (0.10 — 1.28155(0.2)) = 42.7470 Wie use equation (23.6) to calculate the expected value ofthe stock given tha it less than 3902822, We will sneha calculate dj and dy but you can save work if yourelie thal by the definition of d, N(-dz) = 01 ‘n(50/42.76470) + 0.12 + 0.5(0.22) ————— 48155 — 0.2 = 1.28155 06923 Ni) =01 506" (0.06923) Ti ea wc i aca ici er om peso esa 13s [S; | $1 > 72.7755] = = s.0110 28185 59.0082 TVR ofthe payott is east enerQUIZ SOLUTIONS FOR LESSON 23 3a Quiz Solutions 234, The median of $,/S0 ise and (a5 0.50%) (012 -003-a.5(0.4"))t = 0.018 ‘Tomake the median 110, we must have 100 = 110 oot =Inta ¢ = 100in1.1 232. The lognormal paramators are m = 0:1 ~0.5(0258) = 0.06975 and» = 0.25, The prediction interval is obtained by adding and subtracting 1.960 to and from m, or (211.960, m+ 1.960) = (0.06875 ~1.96(0°25) 0.06875 + 1.95(0.25)) = (-0.42125,0.55875) ‘The prediction interval is (902-45, 5024575 23:3, Well use equation (232). (62.81, 97.82) '1n(60/70) + (015-005 -0.540.27))(0.25) 02025 = Ni-1.ss = [a8] NG) 23.4, We use formula (23.10), Notice that K 0 £0 In(S9/K) =0. 7 _ 0+ 050.22) a 02 ds 06-02-04 Nida) 0465642 ‘The expected payor is E[max(0,$:~K)] 100° (0.72575) ~ 100065542) atstedy nat atin 2 pring Peesrtcrrest