Hello

I’ve simulated an AR(3) process: X_t = 5 + 0.4X_t-1 - 0.1X_t-2 + 0.6X_t-3  + U_t

The PACF gives me only for lag 1 a statistically significant value.

Shouldn’t I get for the first three lags statistically significant values?

If the PACF doesn’t help me to determine the order of the AR process how should I proceed to estimate the process order with gretl?

Thanks

Raul