Page 1 of 2

Near singular matrix

Posted: Tue Jun 15, 2010 7:19 pm
by moirapasseron
Please I need help, I'm trying to make a Least Squares analysis but I have trouble when entering my equation because i get the message "Near singular matrix".

the equation is:

c(1) c(2) c(3) c(1)*0.0128*0.507*capital^(0.507-1)-c(2)*(descuento+0.05)+c(3)*(efectividad-(0.507/capital))*consumo

And i have attached my eviews file, please someone knows what's wrong? i need that for my thesis urgently. Thanks

Re: Near singular matrix

Posted: Tue Jun 15, 2010 7:23 pm
by startz
Please I need help, I'm trying to make a Least Squares analysis but I have trouble when entering my equation because i get the message "Near singular matrix".

the equation is:

c(1) c(2) c(3) c(1)*0.0128*0.507*capital^(0.507-1)-c(2)*(descuento+0.05)+c(3)*(efectividad-(0.507/capital))*consumo

And i have attached my eviews file, please someone knows what's wrong? i need that for my thesis urgently. Thanks
The leading

Code: Select all

c(1) c(2) c(3)
probably doesn't belong there.

Re: Near singular matrix

Posted: Tue Jun 15, 2010 7:49 pm
by moirapasseron
OK, i've erased c(1) c(2) c(3) from the equation and the error message dessapeared.
But now the least squares method only returns 0 and 1 values for c(1) c(2) c(3)...what 'im still doing wrong?
thanks and sorry, i'm new in eviews.

Re: Near singular matrix

Posted: Tue Jun 15, 2010 7:50 pm
by moirapasseron
OK, i've erased c(1) c(2) c(3) from the equation and the error message dessapeared.
But now the least squares method only returns 0 and 1 values for c(1) c(2) c(3)...what 'im still doing wrong?
thanks and sorry, i'm new in eviews.

Please I need help, I'm trying to make a Least Squares analysis but I have trouble when entering my equation because i get the message "Near singular matrix".

the equation is:

c(1) c(2) c(3) c(1)*0.0128*0.507*capital^(0.507-1)-c(2)*(descuento+0.05)+c(3)*(efectividad-(0.507/capital))*consumo

And i have attached my eviews file, please someone knows what's wrong? i need that for my thesis urgently. Thanks
The leading

Code: Select all

c(1) c(2) c(3)
probably doesn't belong there.

Re: Near singular matrix

Posted: Tue Jun 15, 2010 9:44 pm
by moirapasseron
ah and i'm using eviews 6, please help!

Re: Near singular matrix

Posted: Tue Jun 15, 2010 9:54 pm
by startz
Try running least squares by listing your dependent variable followed by your independent variables (separated by spaces).

Re: Near singular matrix

Posted: Tue Jun 15, 2010 10:02 pm
by moirapasseron
Try running least squares by listing your dependent variable followed by your independent variables (separated by spaces).

Ok so I have:

m n p v c(1)*0.0128*0.507*m^(0.507-1)-c(2)*(p+0.05)+c(3)*(v-(0.507/m))*n (In the equation estimation)

where m n p v are my independent variables, but the dependent variable i dont know how to distinguishe it ¿? sorry again and thanks for your time!

Near singular matrix

Posted: Tue Jun 15, 2010 10:11 pm
by EViews Gareth
I think, perhaps, you need to figure out your equation mathematically before you can figure out how to estimate it in EViews.

What exactly is the equation you're estimating?

Re: Near singular matrix

Posted: Tue Jun 15, 2010 10:34 pm
by moirapasseron
I think, perhaps, you need to figure out your equation mathematically before you can figure out how to estimate it in EViews.

What exactly is the equation you're estimating?

I'm studying a economic growth model with enviromental externalities involved. The model equation is the following:

gamma = A*B*K^(B-1) - (rho+delta) + v (phi - (B/k))*c

So I have time series from 1970-2000 for k, c, rho and phi (wich I already imported from excel).
For the rest of variables: A, B and delta I have constants values assigned (A=0.0128; B=0.507; delta=0.05)
Finnaly v is the variable that i need to estimate, because I need to define if it is significant for the model or not.

So the equation i've wrote in eviews is:

C(1)*0.0128*0.507*K^(0.507-1)-C(2)*(rho+0.05)+C(3)*(V-(0.507/k))*cons

When i define this equation and run the Least Squares, I only obtain 0 and 1 in the ESTIMATION OUTPUT, what is it that i'm doing wrong? Thanks a lot for your time i really appreciate it.

Re: Near singular matrix

Posted: Tue Jun 15, 2010 10:55 pm
by moirapasseron
I think, perhaps, you need to figure out your equation mathematically before you can figure out how to estimate it in EViews.

What exactly is the equation you're estimating?

Here I uploaded my file, the final equation is: C(1)*0.0128*0.507*K^(0.507-1)-C(2)*(P+0.05)+C(3)*(PHI-(0.507/K))*CONS
Thanks again gareth.

Re: Near singular matrix

Posted: Wed Jun 16, 2010 6:48 am
by startz
I think, perhaps, you need to figure out your equation mathematically before you can figure out how to estimate it in EViews.

What exactly is the equation you're estimating?

I'm studying a economic growth model with enviromental externalities involved. The model equation is the following:

gamma = A*B*K^(B-1) - (rho+delta) + v (phi - (B/k))*c

So I have time series from 1970-2000 for k, c, rho and phi (wich I already imported from excel).
For the rest of variables: A, B and delta I have constants values assigned (A=0.0128; B=0.507; delta=0.05)
Finnaly v is the variable that i need to estimate, because I need to define if it is significant for the model or not.

So the equation i've wrote in eviews is:

C(1)*0.0128*0.507*K^(0.507-1)-C(2)*(rho+0.05)+C(3)*(V-(0.507/k))*cons

When i define this equation and run the Least Squares, I only obtain 0 and 1 in the ESTIMATION OUTPUT, what is it that i'm doing wrong? Thanks a lot for your time i really appreciate it.
Where do you get the values of gamma?

Re: Near singular matrix

Posted: Wed Jun 16, 2010 7:30 am
by moirapasseron
I think, perhaps, you need to figure out your equation mathematically before you can figure out how to estimate it in EViews.

What exactly is the equation you're estimating?

I'm studying a economic growth model with enviromental externalities involved. The model equation is the following:

gamma = A*B*K^(B-1) - (rho+delta) + v (phi - (B/k))*c

So I have time series from 1970-2000 for k, c, rho and phi (wich I already imported from excel).
For the rest of variables: A, B and delta I have constants values assigned (A=0.0128; B=0.507; delta=0.05)
Finnaly v is the variable that i need to estimate, because I need to define if it is significant for the model or not.

So the equation i've wrote in eviews is:

C(1)*0.0128*0.507*K^(0.507-1)-C(2)*(rho+0.05)+C(3)*(V-(0.507/k))*cons

When i define this equation and run the Least Squares, I only obtain 0 and 1 in the ESTIMATION OUTPUT, what is it that i'm doing wrong? Thanks a lot for your time i really appreciate it.
Where do you get the values of gamma?


The values of gamma are given because:
A = rate of available technology (1,28%)
B = capital output share (0,507)
delta = depreciation rate (5%)
All this numbers are for a specific economy wich my model is based on.

To clarify, the variables of the model "gamma = A*B*K^(B-1) - (rho+delta) + v (phi - (B/k))*c" are:
gamma = economic growth path
A, B, delta defined above.
k = gross capital stock
rho = discount rate
v = pollution sensitivity
phi = effectiveness of capital on innovation
c = consumption

Re: Near singular matrix

Posted: Wed Jun 16, 2010 8:11 am
by startz

I'm studying a economic growth model with enviromental externalities involved. The model equation is the following:

gamma = A*B*K^(B-1) - (rho+delta) + v (phi - (B/k))*c

So I have time series from 1970-2000 for k, c, rho and phi (wich I already imported from excel).
For the rest of variables: A, B and delta I have constants values assigned (A=0.0128; B=0.507; delta=0.05)
Finnaly v is the variable that i need to estimate, because I need to define if it is significant for the model or not.

So the equation i've wrote in eviews is:

C(1)*0.0128*0.507*K^(0.507-1)-C(2)*(rho+0.05)+C(3)*(V-(0.507/k))*cons

When i define this equation and run the Least Squares, I only obtain 0 and 1 in the ESTIMATION OUTPUT, what is it that i'm doing wrong? Thanks a lot for your time i really appreciate it.
Where do you get the values of gamma?


The values of gamma are given because:
A = rate of available technology (1,28%)
B = capital output share (0,507)
delta = depreciation rate (5%)
All this numbers are for a specific economy wich my model is based on.

To clarify, the variables of the model "gamma = A*B*K^(B-1) - (rho+delta) + v (phi - (B/k))*c" are:
gamma = economic growth path
A, B, delta defined above.
k = gross capital stock
rho = discount rate
v = pollution sensitivity
phi = effectiveness of capital on innovation
c = consumption
But you say you're trying to estimate v. You don't know gamma unless you know v. So this is circular.

Re: Near singular matrix

Posted: Wed Jun 16, 2010 8:20 am
by moirapasseron
The values of gamma are given because:
A = rate of available technology (1,28%)
B = capital output share (0,507)
delta = depreciation rate (5%)
All this numbers are for a specific economy wich my model is based on.

To clarify, the variables of the model "gamma = A*B*K^(B-1) - (rho+delta) + v (phi - (B/k))*c" are:
gamma = economic growth path
A, B, delta defined above.
k = gross capital stock
rho = discount rate
v = pollution sensitivity
phi = effectiveness of capital on innovation
c = consumption[/quote]

But you say you're trying to estimate v. You don't know gamma unless you know v. So this is circular.[/quote]


yes I don't know gamma, i just assigned known values to some variables (A, B and delta) in gamma.
Or you say I have to eliminate those values for this to work?
thanks again for your attention.

Re: Near singular matrix

Posted: Wed Jun 16, 2010 8:33 am
by startz
To oversimplify, you have two unknowns (v and gamma) and only one equation. There's no way to get a meaningful solution. My recommendation would be to rethink what you're trying to estimate from the beginning.