Dear all, I have the following sas data set generated as bellow.

proc iml; start garch(n,u,alpha0,alpha1,tau,seed) global(yt,y_t,vt,v_t); = j(n+1,1,1); = j(n+1,1,u); = j(n+1,1,u); do t = 2 to n+1; v[t] = alpha0 + alpha1*(y[t-1]-u[t-1])**2 + tau*v[t-1]; u[t] = u[t-1] + sqrt(fator*v[t])*normal(seed); y[t] = u[t] + sqrt(v[t])*normal(seed); end; yt = v[2:(n+1)]; yt = y[2:(n+1)]; y_t = v[1:(n+0)]; y_t = y[1:(n+0)]; create garch var{yt y_t vt v_t}; append; finish garch;

call garch(500,10,0.2,0.2,0.6,0.000,123456);

when I run de proc nlp in this way:

proc nlp data = garch_11 phes pcov tech=trureg; max logf; parms mu = 10, a0= 0.2, a1 = 0.2, tau = 0.6; logf = -0.5E+0*log(a0+a1*(y_t-mu)**2+tau*v_t)-0.5E+0*((yt-mu) **2/(a0+a1*(y_t-mu)**2+tau*v_t)); run;

I have:

U 9.9367281 0 0.193814 1 0.360402 AU 0.532892

and when a run de proc autoreg

proc autoreg data = garch; model yt= / garch=(q=1,p=1) maxit=500 method=ml dist=normal initial=(10,0.2,0.2,0.6); run;

Intercept 9.929473 (mu) ARCH0 0.255313 (a0) ARCH1 0.336761 (a1) GARCH1 0.456073 (tau)

My question: How can I get the same result in both proc? Whats's wrong in the proc nlp.

If anyone can help me i'd GREATLY appreciate it

jOE CALLDER

______________________________________________________ Get Your Private, Free Email at http://www.hotmail.com