Estimation of Stochastic Volatility
Posted: Mon Aug 27, 2012 4:20 amHi,
I am trying to estimate an AR(1) process with stochastic volatility:
z = rho_z*z(-1) + exp(sig_z)*u_z;
sig_z = (1-rho_s)*sigbar + rho_s*sig_z(-1) + eta*u_s;
I have data on z and I would like to use a non-linear estimation using the particle filter. However when I call the estimate() command with the option order=2, I get the following error:
??? Error using ==> stochastic_solvers at 156
2nd and 3rd order approximation not implemented for purely backward models
Any idea what is going on? .... Thanks!
I am trying to estimate an AR(1) process with stochastic volatility:
z = rho_z*z(-1) + exp(sig_z)*u_z;
sig_z = (1-rho_s)*sigbar + rho_s*sig_z(-1) + eta*u_s;
I have data on z and I would like to use a non-linear estimation using the particle filter. However when I call the estimate() command with the option order=2, I get the following error:
??? Error using ==> stochastic_solvers at 156
2nd and 3rd order approximation not implemented for purely backward models
Any idea what is going on? .... Thanks!