1

How to prove second moment of a quadratic form where $Z$ has normal distribution with mean zero and covariance matrix identical?

Richard Hardy
  • 54,375
  • 10
  • 95
  • 219
Brandon Quiñones
  • 11
  • 2
  • Also, see https://stats.stackexchange.com/questions/48066/expected-value-of-quadratic-form – StatsStudent Jun 12 '19 at 02:42
  • In the one-variable case with $T=(t_{11}),$ you are asking us to show $$t_{11}= t_{11}E[Z^2] =E[t_{11}Z^2] = E[Z^\prime T Z] =\operatorname{tr}^2(T) + \operatorname{tr}(T)^2 = t_{11}^2+t_{11}^2 = 2t_{11}^2.$$ – whuber Jun 12 '19 at 15:09

1 Answers1

1

Suppose $A\in\mathbb R^{m\times n}$ and $B\in\mathbb R^{n\times m}.$ Then $\operatorname{tr}(AB) = \operatorname{tr}(BA).$ The proof of that is routine.

So we have $Z\sim N_n(0, I_n)$ and $T\in\mathbb R^{n\times n}.$ Then \begin{align} & \operatorname E(Z'TZ) = \operatorname E(\operatorname{tr}(Z'TZ)) = \operatorname E(\operatorname{tr}(TZZ')) \\[8pt] = {} & \operatorname{tr}(\operatorname E(TZZ')) \quad \text{since tr is linear} \\[8pt] = {} & \operatorname{tr}(T\operatorname E(ZZ')) \quad \text{since $T$ is constant} \\[8pt] = {} & \operatorname{tr}(T I_n) = \operatorname{tr}(T). \end{align} So it appears that your proposed identity is mistaken

Michael Hardy
  • 7,094
  • 1
  • 20
  • 38