F-Testing ; constant returns to scale

1.3k Views Asked by At

$lnQ=1.37+0.632lnK+0.452lnL$

(0.257). (0.219)

$cov(bk,bl)=0.055, R^2=0.98$ $H_0: bk+bl=1$

How can I proceed f-test even though I can't find df and RSS?

1

There are 1 best solutions below

0
On

I would apply the t-test.

$H_0: \beta_k+\beta_l=1$

$H_1: \beta_k+\beta_l\neq 1$

$T=\Large{\frac{\hat \beta_k+\hat \beta_l-1}{\sqrt{var{(\hat \beta_k)} +var{(\hat \beta_l)} -cov(\hat \beta_k,\hat \beta_l)}}}$

You compare the value of T with the value of $t_{(n-3,1-\frac{\alpha}{2})}$. n is the number of observations and $\alpha$ is the significance level.

If $|T|<t$, then you do not reject the Null-Hypothesis.