Page 1 of 1

Estimation problem.

Posted: Wed May 30, 2012 7:46 am
by jayjalvi
I am trying to estimate log(TC) = β1 + β2 log(Q) + β3 log(p1) + β4 log(p2) β5 log(p3) +e

Where p1 to p3 are factors prices for inputs of a firm being wage, capital and fuel respectively).

I am trying to Estimate the equation by the least-squares method but I havnt done log equations before. Any help would be highly appriciated. Also how do we plot squared of residual from the said equation again log(Q) with squared residual on y axis?

Thanks a lot in advance. I am still learning eviews.

FIle is attached.

Re: Estimation problem.

Posted: Wed May 30, 2012 7:55 am
by EViews Gareth
Just estimate it in the same way you would estimate a non-log equation, but rather than typing in your variables type in log(variable) instead.

Re: Estimation problem.

Posted: Wed May 30, 2012 8:01 am
by jayjalvi
log(tc) c log(Q) log(wage) log(capital) log(fuel)

I tried this and it came up with an error saying insuffiecnt number of observations and after several attempts it came up with this.

Dependent Variable: LOG(TC)
Method: Least Squares
Date: 05/31/12 Time: 01:00
Sample: 1 145
Included observations: 145

Variable Coefficient Std. Error t-Statistic Prob.

C -3.526503 1.774367 -1.987471 0.0016
LOG(Q) 0.720394 0.017466 41.24448 NA
LOG(WAGE) 0.436341 0.291048 1.499209 0.1361
LOG(CAPITAL) -0.219888 0.339429 -0.647819 0.5182
LOG(FUEL) 0.426517 0.100369 4.249483 0.0000

S.E. of regression 0.392356 Akaike info criterion 1.000578
Sum squared resid 21.55201 Schwarz criterion 1.103224
Log likelihood -67.54189 Hannan-Quinn criter. 1.042286
Durbin-Watson stat 1.013062

Re: Estimation problem.

Posted: Wed May 30, 2012 8:02 am
by jayjalvi
What I am not sure about it now is that if the output is correct after that error that I got. :(

Thanks a lot for your reply it atleast got me till here :)

Could you please help me with my 2nd concern as well? ABout plotting the squared residuals?

Re: Estimation problem.

Posted: Wed May 30, 2012 8:19 am
by EViews Gareth
Make a series of squared residuals, put it in a group with the other variables you want, then make a graph out of them.

Re: Estimation problem.

Posted: Wed May 30, 2012 8:34 am
by jayjalvi
I already have a series called resid in my workfile and I want to plot it again log(Q) . How do I go about with Sqyared of Residual using the series resid? SOrry for being a pain but I am just trying to get my head around it. I don't want to look annoying just that I can't understand. COuld you please, please explain it a little more.

I shall be very thnakful!

Estimation problem.

Posted: Wed May 30, 2012 9:42 am
by EViews Gareth
Series residsq = resid^2