Given the following stochastic process:

$$Y_t = t X$$

where: $E(X)=0$ and $VAR(X) = 1 $.

Is it correct to obtain the autocovariance in the following way?

$$COV(Y_t, Y_s) = E [Y_tY_s] ~~~~~~~~~~~~~~~~~~~\text{since, $E[X_t] = 0$.}$$

$$E [Y_tY_s]=E[tXsX] = tsE[X^2] = tsVAR(X) = ts$$

Others say me that the correct way is:

$$E [Y_tY_s]=E[tXsX] = tE[X]sE[X] = 0$$

Which one is correct?


Your way is correct. \begin{eqnarray} Cov(Y_t,Y_s) &=& E[Y_t\cdot Y_s]-E[Y_t]\cdot E[Y_s] \\ &=& E[tX\cdot sX]-E[tX]\cdot E[sX] \\ &=& tsE[X^2]-ts(E[X])^2 \\ &=& tsE[X^2] \\ &=& ts(Var[X]-E[X]^2) \\ &=& ts(1-0^2) \\ &=& ts \end{eqnarray}

  • $\begingroup$ Perhaps something to note too is that this process is not stationary, and an auto-covariance function that depends only on the lag (i.e. $t-s$) does not exist. $\endgroup$ Jan 26 '17 at 2:43

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.