Solving a 3 equation system
Posted: Mon May 18, 2015 6:25 am
Hello!
I am trying to estimate a system of 3 equations - a production function and 2 first order conditions - and was hoping for some help! I am not really sure how to write the equations (some of the variables are non-linear and some just seem to be constants) for EViews and I would really appreciate some help writing the equation out. I will be using SUR to estimate the system.
The equations are in the attachment, and I have attempted to write them them for an EViews system below:
log(wL/pY) = c(1) + c(2)*log(1-pi_0) + c(3)*log((Y/Y_0)/(L/L_0)) - c(3)*log(c(4)) - c(3) * (c(5)/c(6))*(t^(c(6)) - t_0)
log(qK/pY) = c(7) + c(8)*log(pi_0) + c(3)*log((Y/Y_0)/(K/K_0)) - c(3)*log(c(4)) - c(3)*(c(9)/c(10))*(t^(c(10)) - t_0)
log(Y/L) = c(11) + log(c(4)*(Y_0/N_0)) + (c(5)/c(6))(t^(c(6))-t_0) - c(3)*log(pi_0*e^(c(3)*(c(5)/c(6)(t^(c(6))-t_0) - c(3)*(c(9)/c(10))(t^c(10)-t_0) * ((K/K_0)/(L/L_0)) + *(1-pi_0))
c(1) c(7) c(11) are the constants for the equations
All the parameters with a _0 after them are constant values, the rest are time series
I am concerned with the coefficients such as c(3)*log(c(4)) which is just a constant term, so wouldnt this just fold into the first constant term? This is also the case with c(2)*log(1-pi_0) , as pi_0 is a constant.
The parameters (c(5)/c(6))(t^(c(6))-t_0) is also difficult, given the number of coefficients in this term, would work in Eviews?
Any help would be greatly appreciated!
I am trying to estimate a system of 3 equations - a production function and 2 first order conditions - and was hoping for some help! I am not really sure how to write the equations (some of the variables are non-linear and some just seem to be constants) for EViews and I would really appreciate some help writing the equation out. I will be using SUR to estimate the system.
The equations are in the attachment, and I have attempted to write them them for an EViews system below:
log(wL/pY) = c(1) + c(2)*log(1-pi_0) + c(3)*log((Y/Y_0)/(L/L_0)) - c(3)*log(c(4)) - c(3) * (c(5)/c(6))*(t^(c(6)) - t_0)
log(qK/pY) = c(7) + c(8)*log(pi_0) + c(3)*log((Y/Y_0)/(K/K_0)) - c(3)*log(c(4)) - c(3)*(c(9)/c(10))*(t^(c(10)) - t_0)
log(Y/L) = c(11) + log(c(4)*(Y_0/N_0)) + (c(5)/c(6))(t^(c(6))-t_0) - c(3)*log(pi_0*e^(c(3)*(c(5)/c(6)(t^(c(6))-t_0) - c(3)*(c(9)/c(10))(t^c(10)-t_0) * ((K/K_0)/(L/L_0)) + *(1-pi_0))
c(1) c(7) c(11) are the constants for the equations
All the parameters with a _0 after them are constant values, the rest are time series
I am concerned with the coefficients such as c(3)*log(c(4)) which is just a constant term, so wouldnt this just fold into the first constant term? This is also the case with c(2)*log(1-pi_0) , as pi_0 is a constant.
The parameters (c(5)/c(6))(t^(c(6))-t_0) is also difficult, given the number of coefficients in this term, would work in Eviews?
Any help would be greatly appreciated!