Permanent and Transitory Movements in Output and the Uneml.
Posted: Sat Oct 17, 2009 3:31 pm
Hi all,
i wish to replicate Sinclair’s paper (The Relationships between Permanent and Transitory Movements in U.S. Output and the Unemployment Rate, 2009) using eviews. My model as fallows:
param C(6) 4.000173 c(8) 3.646685 c(14) 1.267028 c(1) 0.4 c(2) 0.450644 c(16) 0.208434 c(7) 1.4 c(9) -0.179 c(15) -0.67 c(4) -0.919887 c(17) 0.8421129 c(10) 0.591248 c(11) -0.2 c(12) 0.5 c(13) -0.4
@signal gdp=sv1+sv2
@signal ur=sv3+sv4
'define state equation
@state sv1=c(1)+sv1(-1)+[ename = e1, var = exp(c(6))]
@state sv3=sv3(-1)+[ename = e2, var = exp(c(7))]
@state sv2=c(2)*sv2(-1)+c(16)*sv5(-1)+[ename = e3, var = exp(c(8))]
@state sv5=sv2(-1)
@state sv4=c(4)*sv4(-1)+c(17)*sv6(-1)+[ename = e4, var = exp(c(9))]
@state sv6=sv4(-1)
'define cross covariances
@evar cov(e1 ,e2)=c(10)
@evar cov(e1,e4)=c(11)
@evar cov(e2 ,e3)=c(12)
@evar cov(e3 ,e4)=c(13)
'define within covariances
@evar cov(e1,e3)=c(14)
@evar cov(e2,e4)=c(15)
Gdp: gross domistic product
Ur:unemployment rate
Sv1=permanent com. of gdp
Sv2=trans. com. of gdp
Sv3=permanent com. of ur
Sv4=trans.com of ur
As in the paper, i have not included a drift term in the sv3, allowing cross and within correlations of errors between both state and signal equations.
But output is different from the one presented in the paper. An suggestions?
i've included the data file.
i wish to replicate Sinclair’s paper (The Relationships between Permanent and Transitory Movements in U.S. Output and the Unemployment Rate, 2009) using eviews. My model as fallows:
param C(6) 4.000173 c(8) 3.646685 c(14) 1.267028 c(1) 0.4 c(2) 0.450644 c(16) 0.208434 c(7) 1.4 c(9) -0.179 c(15) -0.67 c(4) -0.919887 c(17) 0.8421129 c(10) 0.591248 c(11) -0.2 c(12) 0.5 c(13) -0.4
@signal gdp=sv1+sv2
@signal ur=sv3+sv4
'define state equation
@state sv1=c(1)+sv1(-1)+[ename = e1, var = exp(c(6))]
@state sv3=sv3(-1)+[ename = e2, var = exp(c(7))]
@state sv2=c(2)*sv2(-1)+c(16)*sv5(-1)+[ename = e3, var = exp(c(8))]
@state sv5=sv2(-1)
@state sv4=c(4)*sv4(-1)+c(17)*sv6(-1)+[ename = e4, var = exp(c(9))]
@state sv6=sv4(-1)
'define cross covariances
@evar cov(e1 ,e2)=c(10)
@evar cov(e1,e4)=c(11)
@evar cov(e2 ,e3)=c(12)
@evar cov(e3 ,e4)=c(13)
'define within covariances
@evar cov(e1,e3)=c(14)
@evar cov(e2,e4)=c(15)
Gdp: gross domistic product
Ur:unemployment rate
Sv1=permanent com. of gdp
Sv2=trans. com. of gdp
Sv3=permanent com. of ur
Sv4=trans.com of ur
As in the paper, i have not included a drift term in the sv3, allowing cross and within correlations of errors between both state and signal equations.
But output is different from the one presented in the paper. An suggestions?
i've included the data file.