Results for option pricing

Transcription

1 Results for option pricing [o,v,b]=optimal(rand(1, Estimators = o = % best linear combination (true value= v = e-005 %variance per uniform input b =

2 Efficiency of Optimal Linear Combination Efficiency gain based on number of uniform random numbers / or about 40,000. owever, one uniform generates 5 estimators requiring 10 function evaluations. Efficiency based on function evaluations approx 4,000 A simulation using 500,000 uniform random numbers ; 13 seconds on Pentium IV (.4 Ghz equivalent to twenty billion simulations by crude Monte Carlo.

3 Interpreting the coefficients b. Dropping estimators. Variance of the mean of 100,000 is 1.18 Standard error is around Some weights are negative, (e.g. on Y1 some more than 1 (on Y, some approximately 0 (could they be dropped? For example if we drop variance increases to Y 3 about then

4 Tips If you are simulating to generate resuts for a more complicated model (e.g. asian option, non-normal distribution etc use a simple model (European option, normal distribution etc as contol variate. Use simulation to estimate the difference (assuming you know the result for the control Allow the uniform variates as input parameters. This facilitates variance reduction without changing the program. Try a variety of variance reduction techniques (5-10 including some with second-difference like expressions. Your best estimator is usually an optimal linear combination. Only combine antithetic random numbers additively.

5 Black-Scholes price in R blsprice=function(so,strike,r,t,sigma,div{ d1<-(log(so/strike+(rdiv+(sigma^/*t/(sigma*sqrt(t d<-d1-sigma*sqrt(t call<-so*exp(-div*t*pnorm(d1-exp(- r*t*strike*pnorm(d put=call-so+strike*exp(-r*t c(call,put}

6 Useful URL s (numerical recipies (R library

7 Simulating Survivorship bias and the maxima of Brownian Motion Examples in Biostatistics: Sequential tests for a mean. e.g. For testing hypothesis 0 : µ 0 : µ < 1 Reject 0 i. e.min{ B( t + c 0 as soon as B( t 0 + c t 1 < 0} < c 0 c t. 1

8 Modeling highs

9 Brownian Motion Brownian motion dx ( t = µ dt + σ dw ( t X ( t ~ N ( µ t, σ t

10 We typically observe processes Motion Brownian correlated are ( } 0 ; ( max{ } 0 ; ( min{ ( (0 t X T t t X T t t X L T X C X O i i i i i i i i i < < = < < = = =

11 Acceptance-rejection generating from f(x Suppose we generate the Close C by Acceptance-Rejection using its pdf f(x

12 We can use the same picture to simulate jointly (,C for BM

13 Similarly if C is discrete

14 Exponential Statistics Theorem : For Brownian Motion, Z Z L = ( O( C ~ exp( σ T / = ( L O( L independent of (O,C. Proof : C ~ exp( σ T /

15 Random Walk and reflection P( m C = u = f 1 (m u f ( u m max(0, u m > u

16 Substitute observed values in survivor function in Normal case. is exp( ( Z is exp(1 ( ln( Therefore } ( exp{ } ( ( 1 exp{ ( ( But U Uniform[0,1]. observed values.therefore it is the survivor function for given C evaluated at its is ( ( σ σ σ σ T C T C U T C C C T C f C f C f C f U = = = = = =

17 Estimating volatility based on igh and Low Both Z an estimator of volatility or variance. Alternative to sample variance of increments e.g. average(c- O average( Z and Z L + Z (and average provide L has about five times the efficiency One week high and low equivalent to daily data. Also unbiased regardless of drift in BM.

18 Cumulative Variance 001: US/CAD$ Exchange rate

19 US/CAD$ Exchange rate Volatility igh vol begins Aug 7 98

20 Evident from graph of exchange rate?

21 Moving average Volatility Russian debt crisis

22 Are Z and ZL exponential distributed? US/CAD$ exchange & Dow Jones To transform from Geometric to Brownian Motion, Z Z L = ln( = ln( L / / O O ( / C ln( L / C

23 In practice, measures of volatility may differ. e.g. DJA Intra-day volatility (igh-low estimator>>inter-day vol (open-close

24 RISK MANAGEMENT

25 SURVIVORSIP BIAS E.g. Retrospective studies on performance of mutual fund managers. Present market value is conditional on low >0.

26 What happens to the mean of the survivors as the variance survivemovie increases?