A variation of simulated method of moments, fitting a garch model and matching some functions of its coefficients and residuals.

(omega, theta, xi, rho) = (.044, 5.4, .021, -.44)

<t>Let's see. We have a sample of X_i, each a vector, then MLE is argmax \sum Ln p(X_i | parm). Sometimes we only observe one point X, e.g. time series, then it becomes maximizingLn p(X | parm) = Ln p(X[1] ... X[n] | parm) = \sum Ln p(X[j] | X[j-1] ... X[1] , parm) QuoteOriginally posted by: rezaalm...

but that would give, calling parm* the ideal set and parm the one we are estimating\int Ln{p(X | parm)} p(X | pam*) dX ~ \sum Ln { p(X[j] | parm) }Reza, this sounds right. I thought this was MLE, no?

ML estimate minimizes Kullback-Leibler distance to the impirical distribution.

Hi everybody, i wonder if beta distribution is infinitely divisible? Any references on the subject?Thanks.

In other words, you should use copula methodology, probably Gaussian copula in your case. That would allow establishing dependence for arbitrary marginal distributions.

<t>seems that you are looking at a very general optimization problem, i.e. objective function in nonconvex and may have many local minima. In that case you should look into direct search methods. Simulated annealing is a powerful method, genetic algorithms could be applied as well, e.g. tabu search,...

<t>prospero, B_f/B_d has a stochastic component that comes from the interest rate processes. RN derivative dQ_f/dQ_d=(X_t*B_f(t))/(X_0*B_d(t)), and changing the measure from Q_f to to Q is the same as changing the numeraire from B_f to B_d/X. We can add another uncertainty source exogenously, howeve...

<t>Martingale, thanks, that sounds pretty general, i'd need some time to decipher it for my calculus is a little rusty. I'm pondering over the following question. Consider a two country economy described by correlated interest rates processes. Suppose we know risk neutral measures in each country so...

Would Girsanov still hold for a multi-factor model, e.g. with stochastic vol? What is Girsanov in its most general formulation?Any references will be greatly appreciated.

R has a package for multivariate t.

reza, thanks a lot! it is very helpful.

<t>I am looking to estimate state price density under the market measure. An interest model that I'm using does not permit an analytic solution. Any references/ideas on possible functional form for the density? To put it differently, if I know dynamics of a process under both risk neutral(Q) and mar...

Gill,you're right it's not there anymore. Two papers I mentioned were essentially two versions of the same paper. You can get the later one here.