$\mathbf{\text {Show}}$ $\mathbf{E[\hat{\beta _1}]=\beta _1}$

I have already been able to prove

$ \hat{\beta_1}= \frac{\sum_{i=1}^n (x_i-\bar{x})}{\sum_{i=1}^n (x_i- \bar{x})^2}$

I rewrite it as

$ \hat{\beta_1}= \sum^n_{i=1} c_i y_i$

Taking expectation on both sides:

$ E[\hat{\beta_1}]= E[\sum^n_{i=1} c_i y_i]$

$ E[\hat{\beta_1}]= \sum^n_{i=1} c_i E[y_i]$

$ E[\hat{\beta_1}]= \sum^n_{i=1} c_i [\beta_0 + \beta _1 x_i]$

$ E[\hat{\beta_1}]= \beta_0\sum^n_{i=1} c_i + \beta _1 \sum^n_{i=1} c_ix_i$

In order to proceed, I need to show

$\mathbf{\sum^n_{i=1} c_i =0}$ $\mathbf{\text{and}}$ $\mathbf{\sum^n_{i=1} c_ix_i=1}$.

This is where I am getting stucked. There is no duplication since it does mot involve $Ex_i |u_i]$


1 Answer 1


To show that $\sum c_i=0$ and $\sum c_ix_i=1$, it's important to understand that the expression $$ c_i := \frac{x_i-\bar x}{\sum_{i=1}^n (x_i-\bar x)^2}$$ has a numerator that depends on $i$, but the denominator does not. In the denominator the $i$ is a bound variable (it's an index of summation) so the denominator could just as well be written $\sum_{k=1}^n (x_k-\bar x)^2$, or even viewed as a constant, call it $A$, that doesn't depend on $i$. So for laziness we can write $$c_i := \frac{x_i-\bar x}A. $$ Now calculate, using rules of summation:

$$\sum c_i = \sum \left(\frac{x_i-\bar x}A\right)=\frac1A\left(\sum x_i-\sum\bar x\right),\tag1 $$ where all summations are understood to run from $i=1$ to $n$. To see why (1) equals zero, notice that $\sum\bar x$ is summing $n$ copies of the constant $\bar x$, so it equals $n\bar x$. Finally use the definition $\bar x:=\frac1n\sum x_i$ to conclude $\sum c_i=0$.

Here is a tricky way to show $\sum c_ix_i=1$. Calculate $\sum c_i(x_i-\bar x)$ in two ways: $$\sum c_i(x_i-\bar x)=\sum c_ix_i - \bar x\sum c_i=\sum c_ix_i,\tag2$$ where in the last equality we use $\sum c_i=0$. But also: $$\sum c_i(x_i-\bar x)=\sum \left(\frac{x_i-\bar x}{A} \right)(x_i-\bar x)=\frac{\sum(x_i-\bar x)^2}A=1.\tag3$$ Since (2) and (3) are equal, conclude that $\sum c_ix_i=1$.


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.