function [res]=Pbar_ho(phi,X0,sigma,X,r,t,o,x,l,N)%o is number of jumps
%x is the conditioned diffusion correlation
alpha=X(1);sigmapi=X(2);mupi=X(3);lambda=X(4);nu=exp(mupi+sigmapi.^2/2)-1;mui=log(X0)+(r-sigma.^2/2-lambda*nu+phi)*t+o.*mupi+sigma.*sqrt(alpha).*x;sigmai=(1-alpha).*sigma.^2.*t+o.*sigmapi.^2;p=normcdf(-mui./sigmai);res=factorial(N)./(factorial(l).*factorial(N-l)).*(p.^l).*(1-p).^(N-l);end
I am trying to model a conditional portfolio loss distribution for a portfolio with homogenous assets. Above we have conditioned on a browninan motion B_t=x and jump Y_t=o.
My question is, why does 'res' not sum to one for l=1:N? Is it not the binomial distribution?
Best Answer