I fit AR(1)-GARCH(1,1) models to 100 shares return, but at almost all of them the fitted AR(1) model is not stationary (AR{1}=0.99). How I can find a suitable same model? (for example by using the AIC criterion)


p = 1;
q = 0;
m = length(s);
ntrn = floor(0.80*m);

options = optimoptions(@fmincon,'Algorithm','interior-point');

n1 = arrayfun(@(x) {x}, NaN(1,p));
n2 = arrayfun(@(x) {x}, NaN(1,q));
mdl = arima('AR',n1, 'D',D, 'MA',n2, 'Variance',garch(1,1), 'Distribution','T');
  • 1
    $\begingroup$ It is unusual to find large AR coefficients for financial returns. Are you by any chance fitting raw or logarithmic prices rather than logarithmic returns? $\endgroup$ – Richard Hardy Feb 14 at 12:04
  • $\begingroup$ Thanks, I think so $\endgroup$ – Re Ka Feb 15 at 9:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.