Econometrics

Consider the following model: log(Yi) = a+Bx_i + e_i.
If Y_i denotes wages and X_i denotes years of experience, what is the main assumption that allow you interpret OLS estimate of B as ceteris paribus returns to experience?

Men är så osäker... jag tänker att det är ngt i stil med:

The necessary assumption for interpreting B as ceteris paribus effect is that the mean of the error term, e, given years of experience, X_i, should be equal to zero. We can see that E[e|X]=E[e], so E[e]=0 have to happen in order to interpret B as ceteris paribus returns.

Hi!

The model is

$\log Y(x) = a + Bx + e(x)$

and consequently the logarithmic return is

$\log \frac{Y(x+1)}{Y(x)} = B + e(x+1) - e(x).$

The expected log-return is therefore constant, equal to

$\mathbf{E}(\log \frac{Y(x+1)}{Y(x)}) = B,$

assuming that random variables $e(x)$ have zero expected value for all years of experience ( $x$ ).

Albiki

Svara