
\\\begin{align*}
\varepsilon_t &=\rho \varepsilon_{t-1}+u_t \\
&= u_t+\rho u_{t-1}+\rho^2 u_{t-2}+\rho^3 u_{t-3}+\cdots +\rho^{\infty}u_{t-\infty}\\
&=\sum_{s=0}^{\infty}\rho^s u_{t-s} \\
\end{align*}
\\
\begin{align*}
E[\varepsilon_t] &=E[\sum_{s=0}^{\infty}p^s u_{t-s}]\ \\
&=\sum_{s=0}^{\infty}p^sE[u_{t-s}] \\
&=0 \\
\end{align*}


\begin{align*}
Cov(\varepsilon_t, \varepsilon_{t-1}) &=E[\varepsilon_t\varepsilon_{t-1}] \\
&=E[(\rho\varepsilon_{t-1}+u_t)\varepsilon_{t-1}]\\
&=E[\rho \varepsilon_{t-1}^2+u_t\varepsilon_{t-1}] \\
&=\rho Var(\varepsilon_t)+Cov(u_t,\varepsilon_{t-1}) \\
&=\rho Var(\varepsilon_t) +Cov(u_t,\sum_{s=0}^\infty \rho^s u_{t-1-s)}\\
&=\rho Var(\varepsilon_t)+0 \\
&=\frac{\rho\sigma^2_u}{1-\rho^2}
\end{align*}

\begin{align*}
Cov(\varepsilon_t,\varepsilon_{t-2}) &= E[\varepsilon_t \varepsilon_{t-2}] \\
&=E[\sum_{s=0}^{\infty}\rho^s u_{t-s}\sum \rho^su_{t-s-2}] \\
&=E[(u_t+\rho u_{t-1}+\rho^2 u_{t-2}\cdots )(u_{t-2}+\rho u_{t-3}+\cdots )] \\
&=\sum^{\infty}_{s=2} \rho^{2s-2}E[u^2_{t-s}]\\
&=\sigma^2_u \times \frac{\rho^2}{1-\rho^2}\\
\end{align*}
'계량경제학' 카테고리의 다른 글
11.04 수업 정리 (0) | 2021.11.24 |
---|---|
오차항의 표본분산의 일치성 증명 (0) | 2021.11.23 |
d (0) | 2021.11.19 |
11.16 -1 (0) | 2021.11.16 |
11월 11일 자기상관(autocorrelation) (0) | 2021.11.16 |