Initially I am aware of the fact that in order to analyse autocorrelations of residuals of ARMA(p,q) we apply Ljung-Box test and involve chi-square distribution (because autocorrelations are asymptotically normally distributed) with degree of freedom n-p-q. I know the proof of autocorrelations are asymptotically normal but can not derive their sum of square have n-p-q degree of freedom. Could you please provide me with derivation of it?


  • $\begingroup$ In time series, according to ljung-box there is a fact that autocorrelations are normally distributed asymptotically and summation of autocorrelation of errors (such as lag 1, lag 2...) are chi square distribution. Wikipedia explains it. $\endgroup$ – mertcan Mar 30 at 14:20
  • $\begingroup$ I hope my question is explicit, I have not received any significant responses yet. Could you help me? $\endgroup$ – mertcan Mar 31 at 9:42
  • $\begingroup$ Is there anyone who may answer my question? I have not received any responses....... $\endgroup$ – mertcan Apr 4 at 17:33

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.