Limit theorems on a linear explosive stochastic model for time series with moving average error (Q761750)

Let \(\{\) X(t)\(\}\) be a nonstationary (explosive) process generated by \(X(t+q)=b_ 1X(t+q-1)+...+b_ qX(t)+b_ 0+a_ 0\epsilon (t+q)+...+a_{\ell}\epsilon (t-\ell)\), where \(a_ 0=1\) and \(\{\) \(\epsilon\) (t)\(\}\) is a white noise. The authors propose two kinds of estimators for parameters \(b_ 0,...,b_ q\) and prove their consistency and asymptotic normality. They derive also asymptotic tests of fit.