0
$\begingroup$

Define the linear regression model: $$ Y_i = \beta_0 + \beta_1 X_i + \varepsilon_i, \quad \quad i = 1,\dots,n. $$ Let $\hat \beta_0$ and $\hat \beta_1$ be the estimates of $\beta_0$ and $\beta_1$ when we solve the regression model with least squares. $\hat \beta_0$ and $\hat \beta_1$ are dependent on the $X_i$'s and $Y_i$'s since they were computed using them.

I am wondering what is the expected value of $\hat \beta_1 - \beta_1$ multipled by one of the $X_i$'s, say $X_1$ for example. That is, what is $E[(\hat \beta_1 - \beta_1)X_1]$?

$\endgroup$

1 Answer 1

1
$\begingroup$

$$ \mathsf{E}\!\left[(\hat{\beta}_1-\beta_1)X_j\right]=\mathsf{E}\!\left[\frac{\sum_{i=1}^n(X_i-\bar{X})\varepsilon_iX_j}{\sum_{i=1}^n (X_i-\bar{X})^2}\right]. $$ If $\mathsf{E}[\varepsilon_i\mid X_1,\ldots,X_n]=0$, $$ \mathsf{E}\!\left[(\hat{\beta}_1-\beta_1)X_j\right]=0. $$

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .