Ecotrix Doubt..Gujrati

[The Villain]

This is a snapshot of pg 199 gujrati..
The place where i have put a pointer ...does it mean beta 2 hat cannot be Y bar by X bar...pls help someone!!!!

[Akshay Jain]

yes ron this means that B2 hat cannot b Ybar/Xbar.......its the  property of the OLS regression that cov(x,u)=0 and for unbiasedness we use this property.....

[The Villain]

Okay ...Thanxx Akshayy

[The Villain]

pls expalain me this also...pg 206 of gujrati

[Akshay Jain]

@ ron ...this is the concept of partial derivatives that we generally use......
the interpretation of partal derivatives say that what will be the impact of a unit change in 1 explanatory variable on the dependent variable, keeping all the other as constant.
this is what is happening here
the 1st reg is of CM on FLR and the u1i term takes all the effect on CM not explained by the FLR
now we want to take that effect of PGNP on CM that is purely due to change in PGMP and no effect of FLR so we also reg PGNP on FLR to take out any effect on PGNP of FLR ..the u2i is unexplained by FLR
now the u1i and u2i are terms which hav no influence of FLR
if we reg U1i on U2i then we will get the true effect of PGNP on CM by keeping out the effect of FLR from boh the variables.
this is what is called as partial derivative of a multiple linear reg...
the beta coefficients in the MLRM is just like the partial derivative and gives the impact of only 1 variable keeping all other as constant or keeping the effect of other variables out.

[Akshay Jain]

I can make it more clear to you...but m not getting the right words to elaborate this......

[The Villain]

thankuuu so much Akshay..appreciate