EViews.com

Posted: **Wed Mar 23, 2011 4:00 am**

Hi,

I have a question about saving my output in a matrix.
I have a program that regresses 49 industry each year of the sample period, so 49 industries and 41 years. I need the coefficients, st.errors and t-stats. The problem i have is that the matrix should store the first year in column 1 and the second year in column 2 etc. But the problem is that eviews continues in the first column, i have read the programming guide but can't find a way to do this.
this is my program so far.

Code: Select all

' first try 

'preparing to save beta output in table
	matrix(industry.@count,46) betamin
	matrix(industry.@count,46) betaplus
	matrix(industry.@count,46) Tstatbmin
	matrix(industry.@count,46) Tstatbplus
	matrix(industry.@count,46) stderbmin
	matrix(industry.@count,46) stderbplus
	
	!rowcounter=1

'rolling sample (dates)
	for !year= 1969 to 2009

'first year
  	smpl !year !year 

'the different industries ols
	for !i = 1 to industry.@count
	%name = industry.@seriesname (!i)
	equation eq{!year}_{%name}.ls(n)industry(!i) c variables

'getting the results in the table
	betamin(!rowcounter, ??? )= eq{!year}_{%name}.@coefs(2)
	betaplus(!rowcounter, ???)= eq{!year}_{%name}.@coefs(3)
	Tstatbmin(!rowcounter, ??? )= eq{!year}_{%name}.@tstat(2)
	Tstatbplus(!rowcounter, ???)= eq{!year}_{%name}.@tstat(3)	
	stderbmin(!rowcounter, ??? )= eq{!year}_{%name}.@stderrs(2)
	stderbplus(!rowcounter, ???)= eq{!year}_{%name}.@stderrs(3)
'to "roll" to next year
	smpl !year+1 !year+1

'to assure right output is in the corresponding row
	!rowcounter= !rowcounter+1

		
'deleting the equations
	delete eq{!year}_{%name}
next
next

smpl @all

hope you can help me to fill in the question marks

Posted: **Wed Mar 23, 2011 7:55 am**

Based on your current code, probably the easiest thing to do is use a !colcounter variable. Start it at one, and then each time you run through the !year loop, increase it by one.

Posted: **Fri Mar 25, 2011 2:56 am**

Thanks, i used a different way, changed the order of the coding of the results so instead of betamin(!rowcounter,?) I used betamin(!i,!rowcounter)
and changed row counter in !rowcounter= !year-1968.
now it works fine but thanks for your help

EViews.com

Multiple regression:Question about storing output

Multiple regression:Question about storing output

Re: Multiple regression:Question about storing output

Re: Multiple regression:Question about storing output