MethodError: Cannot convert an object of type Nothing to an object of type Int64
#90
Comments
|
I would need the whole model in order to have a look at what might be going on here. |
|
Ok, here is the code for the whole model. Julia codeusing MacroModelling
@model Baseline begin
## 1 FOC labor demand
lvar[0] - mcvar[0] + w[0] - y[0] = 0
##2 FOC capital utilization
(1 - alph) * k[-1] = (1 - alph)*lvar[0] + mcvar[0] - pi[0] + tfp[0] + (-1 + alph - edp)*uvar[0] + (-1 + thetta) * a[0]
## 3 FOC capital demand (definition of dividend)
alph*betta*mcvar[0] + d[0]*(-1 + betta - gy)*KY*muin*muvar + (-(alph*betta) + (1 - betta + gy)*KY*muin*muvar)*pi[0] - alph*betta*uvar[0] + alph*betta*y[0] - alph*betta*k[-1] = 0
## 4 Production function
(1 - alph)*lvar[0] + tfp[0] + alph*uvar[0] - y[0] + (-1 + thetta)*a[-1] + alph*k[-1] = 0
## 5 Marginal utility of consumption
ctil[0]*(-1 - gy + h)*(-1 - gy + betta*h) + c[0]*((1 + gy)^2 + betta*h^2) - (1 + gy)*h*c[-1] - betta*(1 + gy)*h*c[1] = 0
## 6 Euler equation
ctil[1] = ctil[0] - mub[0] - rfvar[0]
## 7 Capital arbitrage condition
ctil[0]*(1 + gy) + (1 + gy)*pi[0] + (-1 - gy)*ctil[1] + (-1 + betta - gy)*d[1] - betta*pi[1] = 0
## 8 Optimal investment (price of capital)
(-1 - betta/(1+gy))*ddac*invs[0] + pi[0] - pinv[0] + ddac*invs[-1] + betta/(1+gy)*ddac*invs[1] = 0
## 9 Capital law of motion
betta*invs[0]*IY - betta*(1 + gy)*k[0]*KY - (1 + betta*(-1 + delta) + gy)*KY*uvar[0] - betta*(-1 + delta)*KY*k[-1] = 0
## 10 Taylor Rule
-mpvar[0] + lvar[0]*phi_y*(-1 + rho_smooth) + phi_m*pivar[0]*(-1 + rho_smooth) + Rnomvar[0] - rho_smooth*Rnomvar[-1] = 0
## 11 Fisher equation
rfvar[0] - Rnomvar[0] + pivar[1] = 0
## 12 Tech LOM
chird[0]*sslabsr^rhoz*varphibar + labsr[0]*rhoz*sslabsr^rhoz*varphibar + (-1 - ga)*z[0] + (1 + ga)*z[-1] = 0
## 13 FOC RD
chird[0] - ctil[0] + labsr[0]*(-1 + rhoz) - ws[0] - z[0] + ctil[1] + jz[1] + z[-1] = 0
## 14 Value of Unadopted Intermediate
-((1 + ga)*jz[0]*JZY*ssa) - a[0]*betta*lam*(1 - obso)*ssz*VAY - (1 + ga)*labsaz[0]*ssa*sslabsaz*WSY - (1 + ga)*ssa*sslabsaz*ws[0]*WSY - ctil[0]*(1 + ga)*ssa*(JZY + sslabsaz*WSY) + lamvar[0]*ssa*(JZY*(1 + ga - betta*(1 - obso)) + (1 + ga)*sslabsaz*WSY) + (-((1 + ga)*JZY*ssa) + betta*lam*(1 - obso)*ssz*VAY - (1 + ga)*ssa*sslabsaz*WSY)*z[0] + (1 + ga)*ssa*(JZY + sslabsaz*WSY)*ctil[1] + ((1 + ga)*JZY*ssa - betta*lam*(1 - obso)*ssz*VAY + (1 + ga)*ssa*sslabsaz*WSY)*jz[1] + betta*lam*(1 - obso)*ssz*VAY*va[1] + (1 + ga)*ssa*(JZY + sslabsaz*WSY)*z[-1] = 0
## 15 Adoption Probability
-lamvar[0] + labsaz[0]*rholambda = 0
## 16 Adopted LOM
-(a[0]*(1 + ga)*ssa) + labsaz[0]*(ga + obso)*rholambda*ssa - (-1 + lam)*(1 - obso)*ssa*a[-1] + lam*(1 - obso)*ssz*z[-1] = 0
## 17 Value of Adopted Intermediate
(-1 - ga)*mcvar[0]*(-1 + muin) + a[0]*betta*muin*muvar*(1 - obso)*VAY + betta*ctil[0]*muin*muvar*(1 - obso)*VAY + (1 + ga)*muin*muvar*va*VAY + (-1 - ga)*(-1 + muin)*y[0] - betta*muin*muvar*(1 - obso)*VAY*a[-1] - betta*muin*muvar*(1 - obso)*VAY*ctil[1] - betta*muin*muvar*(1 - obso)*VAY*va[1] = 0
## 18 FOC adoption
-(a[0]*betta*lam*(1 - obso)*rholambda*ssz*VAY) - ctil[0]*(1 + ga)*ssa*sslabsaz*WSY + (1 + ga)*labsaz[0]*(-1 + rholambda)*ssa*sslabsaz*WSY - (1 + ga)*ssa*sslabsaz*ws[0]*WSY + (betta*lam*(1 - obso)*rholambda*ssz*VAY - (1 + ga)*ssa*sslabsaz*WSY)*z[0] + (1 + ga)*ssa*sslabsaz*WSY*ctil[1] + (-(betta*lam*(1 - obso)*rholambda*ssz*VAY) + (1 + ga)*ssa*sslabsaz*WSY)*jz[1] + betta*lam*(1 - obso)*rholambda*ssz*VAY*va[1] + (1 + ga)*ssa*sslabsaz*WSY*z[-1] = 0
## 19 Market clearing consumption goods
c[0]*CY + g[0]*GY + invs[0]*IY + IY*pinv[0] - y[0] = 0
## 20 Market clearing research labor
labs[0]*sslabs*ssz - labsr[0]*sslabsr*ssz + labsaz[0]*(-sslabs + sslabsr)*ssz + ssa*sslabsaz*a[-1] - ssa*sslabsaz*z[-1] = 0
## % wage setting
(1+kw[0])*w[0]-1/(1+betta)*w[-1]-betta/(1+betta)*w[1]+kw[0]*(ctil[0]-phi*lvar[0])-indw/(1+betta)*pivar[-1]+(1+betta*indw)/(1+betta)*pivar[0]-(betta)/(1+betta)*pivar[1]-muwvar[0] = 0
(1+kw[0])*ws[0]-1/(1+betta)*ws[-1]-betta/(1+betta)*ws[1]+kw[0]*(ctil[0]-phi*labs[0])-indw/(1+betta)*pivar[-1]+(1+betta*indw)/(1+betta)*pivar[0]-(betta)/(1+betta)*pivar[1] = 0
## philips curve
pivar[0]=(indp/(1+betta*indp))*(pivar[-1])+(betta/(1+betta*indp))*(pivar[1])+((1-betta*epsp)*(1-epsp)/(epsp*(1+betta*indp)))*mcvar[0]+mu[0] = 0
## adoption and r&d expenditure
rd_exp[0] = ws[0] + labsr[0]
adopt_exp[0] = ws[0] + labsaz[0] + ssa/(ssz-ssa)*(z[-1]-a[-1]) = 0
## NOTE shocks dynamics
tfp[0] - rhotfp*tfp[-1] - etfp[x]
pinv[0] - rhopk*pinv[-1] - epinv[x]
g[0] - rhog*g[-1] - eg[x]
mub[0] - rhob*mub[-1] - eb[x]
mu[0] - rhomu*mu[-1] - emu[x]
muwvar[0] - rhow*muwvar[-1] - emuw[x]
mpvar[0] - rhomp*mpvar[-1] - empvar[x]
chird[0] - rhochi*chird[-1] - echi[x]
## NOTE measurement equations
Dy[0]=gy*100+(y[0]-y[-1])
Dc[0]=gy*100+(c[0]-c[-1])
Di[0]=gy*100+(invs[0]-invs[-1])+(pinv[0]-pinv[-1])
lvarobs[0]=lvar[0]
Dw[0]=gy*100+w[0]-w[-1]
Rnomobs[0] = Rnomvar[0]+((1-mubss)*(1+gy)/betta-1)*100
pinfobs[0] = pivar[0]+piconst
DrdNSF[0] = gy*4*100+rd_exp[0]-rd_exp[-4]
end
@parameters Baseline begin
# liquidity premium in steady state 50 bps annually
mubss = 0.005/4
delta = 0.02
GY = 0.16
lam = 0.05
muin = 1.18
muvar = 1.1
obso = 0.02
rholambda = 0.925
thetta = 1.35
varphibar = 0.05
rconst = 0.5
piconst = 0.5
constelab = 0.013
## NOTE estimated parameters (latter)
std_eg = 2.5595
std_epinv = 1.2712
std_eb = 0.22225
std_emu = 0.093091
std_etfp = 0.48905
std_empvar = 0.096492
std_emuw = 0.28711
std_echi = 2.2119
rhotfp = 0.95092
rhopk = 0.8999
rhob = 0.92619
rhomp = 0.47126
rhomu = 0.40463
rhog = 0.9681
rhow = 0.27224
rhochi = 0.80216
rho_smooth = 0.82854
phi_m = 1.5745
phi_y = 0.38614
phi = 2.7026
ddac = 5.6386
epsp = 0.9346
epsw = 0.90568
indp = 0.24118
indw = 0.38198
h = 0.48303
edp = 4.0442
rhoz = 0.37076
alph = 0.20068
gyest = 0.45317
bettaest = 0.42122
muwest = 0.15
Rf_zlb = 100*(-1 + (1 + gyest/100)*(bettaest/100+1))
## NOTE: Local-scope variables in dynare
gy = gyest/100
betta = (1 + bettaest/100)^(-1)
ssa = 1
ssR = (1-mubss)*(1+gy)/betta
KY = alph/(muin*muvar*(ssR-1+delta))
IY = KY*(delta+gy)
CY = 1 - GY - IY
ga = -1 + (1 + gy)^((1 - alph)/(-1 + thetta))
ssz = ((1 + ga + (-1 + lam)*(1 - obso))*ssa)/(lam*(1 - obso))
sslabsr = ((ga + obso)/varphibar)^(1/rhoz)
VAY = ((1 + ga)*(-1 + muin))/(muin*muvar*(1 + ga - betta*(1 - obso)))
sslabsaz = ((-1 - ga + betta*(1 - obso))*rholambda*sslabsr^(1 - rhoz))/(betta*(-1 + rholambda)*varphibar)
WSY = -((betta^2*lam*(1 - obso)*(-1 + rholambda)*sslabsr^(-1 + rhoz)*ssz*varphibar*VAY)/((1 + ga)*(1 + ga + betta*(1 - obso)*(-1 + lam - lam*rholambda))*ssa))
JZY = -((betta*lam*(1 - obso)*(-1 + rholambda)*ssz*VAY)/((1 + ga + betta*(1 - obso)*(-1 + lam - lam*rholambda))*ssa))
sslabs = sslabsaz + sslabsr - (ssa*sslabsaz)/ssz
muw = muwest+1
kw = (1-epsw)*(1-beta*epsw)/(epsw*(1+beta))/(1+phi/(1+(1/muw)))
endDynare code (for reference)var c invs Rnomvar w lvar mcvar y k pi uvar a d ctil rfvar pivar z labsr ws jz labsaz lamvar va labs tfp mub pinv mpvar chird g mu muwvar Dy Di pinfobs Dc lvarobs DrdNSF Dw Rnomobs rd_exp adopt_exp;
parameters alpha ddac delta edp GY h lam muin muvar obso phi_y phi_m rholambda rho_smooth rhoz thetta varphibar betaest gyest phi muwest indw indp epsp epsw rhotfp rhopk rhow rhob rhog rhomu rhomp rhochi std_eb std_eg std_emu std_empvar std_epinv std_etfp std_echi std_emuw rconst piconst constelab Rf_zlb mubss;
varexo eb eg etfp epinv emu emuw echi empvar;
set_param_value('alpha',alpha);
set_param_value('ddac',ddac);
set_param_value('delta',delta);
set_param_value('edp',edp);
set_param_value('GY',GY);
set_param_value('h',h);
set_param_value('lam',lam);
set_param_value('muin',muin);
set_param_value('muvar',muvar);
set_param_value('obso',obso);
set_param_value('phi_y',phi_y);
set_param_value('phi_m',phi_m);
set_param_value('rholambda',rholambda);
set_param_value('rho_smooth',rho_smooth);
set_param_value('rhoz',rhoz);
set_param_value('thetta',thetta);
set_param_value('varphibar',varphibar);
set_param_value('betaest',betaest);
set_param_value('gyest',gyest);
set_param_value('phi',phi);
set_param_value('indw',indw);
set_param_value('indp',indp);
set_param_value('epsp',epsp);
set_param_value('epsw',epsw);
set_param_value('rhotfp',rhotfp);
set_param_value('rhopk',rhopk);
set_param_value('rhow',rhow);
set_param_value('rhob',rhob);
set_param_value('rhog',rhog);
set_param_value('rhomu',rhomu);
set_param_value('rhomp',rhomp);
set_param_value('rhochi',rhochi);
set_param_value('std_eb',std_eb);
set_param_value('std_eg',std_eg);
set_param_value('std_emu',std_emu);
set_param_value('std_empvar',std_empvar);
set_param_value('std_epinv',std_epinv);
set_param_value('std_etfp',std_etfp);
set_param_value('std_echi',std_echi);
set_param_value('std_emuw',std_emuw);
set_param_value('rconst',rconst);
set_param_value('piconst',piconst);
set_param_value('constelab',constelab);
set_param_value('muwest',muwest);
set_param_value('mubss',mubss);
Rf_zlb=100*(-1 + (1 + gyest/100)*(betaest/100+1));
model(linear);
#gy = gyest/100;
#beta = (1 + betaest/100)^(-1);
#ssa = 1;
#ssR = (1-mubss)*(1+gy)/beta;
%#KY =(alpha*beta)/((1 + beta*(-1 + delta) + gy)*muin*muvar);
#KY = alpha/(muin*muvar*(ssR-1+delta));
%#IY = (alpha*beta*(delta + gy))/((1 + beta*(-1 + delta) + gy)*muin*muvar);
#IY = KY*(delta+gy);
#CY = 1 - GY - IY;
#ga = -1 + (1 + gy)^((1 - alpha)/(-1 + thetta));
#ssz = ((1 + ga + (-1 + lam)*(1 - obso))*ssa)/(lam*(1 - obso));
#sslabsr = ((ga + obso)/varphibar)^(1/rhoz);
#VAY = ((1 + ga)*(-1 + muin))/(muin*muvar*(1 + ga - beta*(1 - obso)));
#sslabsaz = ((-1 - ga + beta*(1 - obso))*rholambda*sslabsr^(1 - rhoz))/(beta*(-1 + rholambda)*varphibar);
#WSY = -((beta^2*lam*(1 - obso)*(-1 + rholambda)*sslabsr^(-1 + rhoz)*ssz*varphibar*VAY)/((1 + ga)*(1 + ga + beta*(1 - obso)*(-1 + lam - lam*rholambda))*ssa));
#JZY = -((beta*lam*(1 - obso)*(-1 + rholambda)*ssz*VAY)/((1 + ga + beta*(1 - obso)*(-1 + lam - lam*rholambda))*ssa));
#sslabs = sslabsaz + sslabsr - (ssa*sslabsaz)/ssz;
#muw = muwest+1;
%1 FOC labor demand
lvar - mcvar + w - y;
%2 FOC capital utilization
(1 - alpha)*lvar + mcvar - pi + tfp + (-1 + alpha - edp)*uvar + (-1 + thetta)*a(-1) + (-1 + alpha)*k(-1);
%3 FOC capital demand (definition of dividend)
alpha*beta*mcvar + d*(-1 + beta - gy)*KY*muin*muvar + (-(alpha*beta) + (1 - beta + gy)*KY*muin*muvar)*pi - alpha*beta*uvar + alpha*beta*y - alpha*beta*k(-1);
%4 Production function
(1 - alpha)*lvar + tfp + alpha*uvar - y + (-1 + thetta)*a(-1) + alpha*k(-1);
%5 Marginal utility of consumption
ctil*(-1 - gy + h)*(-1 - gy + beta*h) + c*((1 + gy)^2 + beta*h^2) - (1 + gy)*h*c(-1) - beta*(1 + gy)*h*c(+1);
%6 Euler equation
ctil - mub - rfvar - ctil(+1);
%7 Capital arbitrage condition
ctil*(1 + gy) + (1 + gy)*pi + (-1 - gy)*ctil(+1) + (-1 + beta - gy)*d(+1) - beta*pi(+1);
%8 Optimal investment (price of capital)
(-1 - beta/(1+gy))*ddac*invs + pi - pinv + ddac*invs(-1) + beta/(1+gy)*ddac*invs(+1);
%9 Capital law of motion
beta*invs*IY - beta*(1 + gy)*k*KY - (1 + beta*(-1 + delta) + gy)*KY*uvar - beta*(-1 + delta)*KY*k(-1);
%10 Taylor Rule
-mpvar + lvar*phi_y*(-1 + rho_smooth) + phi_m*pivar*(-1 + rho_smooth) + Rnomvar - rho_smooth*Rnomvar(-1);
%11 Fisher equation
rfvar - Rnomvar + pivar(+1);
%12 Tech LOM
chird*sslabsr^rhoz*varphibar + labsr*rhoz*sslabsr^rhoz*varphibar + (-1 - ga)*z + (1 + ga)*z(-1);
%13 FOC RD
chird - ctil + labsr*(-1 + rhoz) - ws - z + ctil(+1) + jz(+1) + z(-1);
%14 Value of Unadopted Intermediate
-((1 + ga)*jz*JZY*ssa) - a*beta*lam*(1 - obso)*ssz*VAY - (1 + ga)*labsaz*ssa*sslabsaz*WSY - (1 + ga)*ssa*sslabsaz*ws*WSY - ctil*(1 + ga)*ssa*(JZY + sslabsaz*WSY) + lamvar*ssa*(JZY*(1 + ga - beta*(1 - obso)) + (1 + ga)*sslabsaz*WSY) + (-((1 + ga)*JZY*ssa) + beta*lam*(1 - obso)*ssz*VAY - (1 + ga)*ssa*sslabsaz*WSY)*z + (1 + ga)*ssa*(JZY + sslabsaz*WSY)*ctil(+1) + ((1 + ga)*JZY*ssa - beta*lam*(1 - obso)*ssz*VAY + (1 + ga)*ssa*sslabsaz*WSY)*jz(+1) + beta*lam*(1 - obso)*ssz*VAY*va(+1) + (1 + ga)*ssa*(JZY + sslabsaz*WSY)*z(-1);
%15 Adoption Probability
-lamvar + labsaz*rholambda;
%16 Adopted LOM
-(a*(1 + ga)*ssa) + labsaz*(ga + obso)*rholambda*ssa - (-1 + lam)*(1 - obso)*ssa*a(-1) + lam*(1 - obso)*ssz*z(-1);
%17 Value of Adopted Intermediate
(-1 - ga)*mcvar*(-1 + muin) + a*beta*muin*muvar*(1 - obso)*VAY + beta*ctil*muin*muvar*(1 - obso)*VAY + (1 + ga)*muin*muvar*va*VAY + (-1 - ga)*(-1 + muin)*y - beta*muin*muvar*(1 - obso)*VAY*a(-1) - beta*muin*muvar*(1 - obso)*VAY*ctil(+1) - beta*muin*muvar*(1 - obso)*VAY*va(+1);
%18 FOC adoption
-(a*beta*lam*(1 - obso)*rholambda*ssz*VAY) - ctil*(1 + ga)*ssa*sslabsaz*WSY + (1 + ga)*labsaz*(-1 + rholambda)*ssa*sslabsaz*WSY - (1 + ga)*ssa*sslabsaz*ws*WSY + (beta*lam*(1 - obso)*rholambda*ssz*VAY - (1 + ga)*ssa*sslabsaz*WSY)*z + (1 + ga)*ssa*sslabsaz*WSY*ctil(+1) + (-(beta*lam*(1 - obso)*rholambda*ssz*VAY) + (1 + ga)*ssa*sslabsaz*WSY)*jz(+1) + beta*lam*(1 - obso)*rholambda*ssz*VAY*va(+1) + (1 + ga)*ssa*sslabsaz*WSY*z(-1);
%19 Market clearing consumption goods
c*CY + g*GY + invs*IY + IY*pinv - y;
%20 Market clearing research labor
labs*sslabs*ssz - labsr*sslabsr*ssz + labsaz*(-sslabs + sslabsr)*ssz + ssa*sslabsaz*a(-1) - ssa*sslabsaz*z(-1);
#kw=(1-epsw)*(1-beta*epsw)/(epsw*(1+beta))/(1+phi/(1+(1/muw)));
% wage setting
(1+kw)*w-1/(1+beta)*w(-1)-beta/(1+beta)*w(+1)+kw*(ctil-phi*lvar)-indw/(1+beta)*pivar(-1)+(1+beta*indw)/(1+beta)*pivar-(beta)/(1+beta)*pivar(+1)-muwvar;
(1+kw)*ws-1/(1+beta)*ws(-1)-beta/(1+beta)*ws(+1)+kw*(ctil-phi*labs)-indw/(1+beta)*pivar(-1)+(1+beta*indw)/(1+beta)*pivar-(beta)/(1+beta)*pivar(+1);
% philips curve
pivar=(indp/(1+beta*indp))*(pivar(-1))+(beta/(1+beta*indp))*(pivar(+1))+((1-beta*epsp)*(1-epsp)/(epsp*(1+beta*indp)))*mcvar+mu;
% adoption and r&d expenditure
rd_exp = ws + labsr;
adopt_exp = ws + labsaz + ssa/(ssz-ssa)*(z(-1)-a(-1));
% shocks dynamics
tfp-rhotfp*tfp(-1)-etfp;
pinv-rhopk*pinv(-1)-epinv;
g-rhog*g(-1)-eg;
mub-rhob*mub(-1)-eb;
mu-rhomu*mu(-1)-emu;
muwvar-rhow*muwvar(-1)-emuw;
mpvar-rhomp*mpvar(-1)-empvar;
chird-rhochi*chird(-1)-echi;
% measurement equations
Dy=gy*100+(y-y(-1));
Dc=gy*100+(c-c(-1));
Di=gy*100+(invs-invs(-1))+(pinv-pinv(-1));
lvarobs=lvar;
Dw=gy*100+w-w(-1);
Rnomobs = Rnomvar+((1-mubss)*(1+gy)/beta-1)*100;
pinfobs=pivar+piconst;
DrdNSF=gy*4*100+rd_exp-rd_exp(-4);
end;
shocks;
var eg; stderr std_eg;
var eb; stderr std_eb;
var emu; stderr std_emu;
var etfp; stderr std_etfp;
var empvar; stderr std_empvar;
var epinv; stderr std_epinv;
var emuw; stderr std_emuw;
var echi; stderr std_echi;
end;
varobs Dw Dc Dy lvarobs Di Rnomobs pinfobs DrdNSF;
stoch_simul(order=1,irf=100,nograph); |
|
The source of the error you got is still unknown to me but essentially the system of equations is not well defined. The error handling in the model and parameter blocks is still poor and I will try to improve the errors so that they are more informative. In any case, you don't need to worry about that if you simply want to translate the model. You can use the function: What you would get then is the following model (which works): @model model_translated_from_dynare begin
lvar[0] - mcvar[0] + w[0] - y[0] = 0
mcvar[0] + lvar[0] * (1 - alpha) - pi[0] + tfp[0] + (alpha - 1 - edp) * uvar[0] + (thetta - 1) * a[-1] + (alpha - 1) * k[-1] = 0
mcvar[0] * alpha * beta + d[0] * (beta - 1 - gy) * KY * muin * muvar + pi[0] * (muvar * muin * KY * (gy + 1 - beta) - alpha * beta) - uvar[0] * alpha * beta + y[0] * alpha * beta - k[-1] * alpha * beta = 0
(thetta - 1) * a[-1] + lvar[0] * (1 - alpha) + tfp[0] + alpha * uvar[0] - y[0] + alpha * k[-1] = 0
ctil[0] * (( - 1) - gy + h) * (( - 1) - gy + beta * h) + c[0] * ((1 + gy) ^ 2 + beta * h ^ 2) - h * (1 + gy) * c[-1] - h * beta * (1 + gy) * c[1] = 0
ctil[0] - mub[0] - rfvar[0] - ctil[1] = 0
ctil[0] * (1 + gy) + pi[0] * (1 + gy) + (( - 1) - gy) * ctil[1] + (beta - 1 - gy) * d[1] - beta * pi[1] = 0
pi[0] + (( - 1) - beta / (1 + gy)) * ddac * invs[0] - pinv[0] + ddac * invs[-1] + beta / (1 + gy) * ddac * invs[1] = 0
beta * invs[0] * IY - KY * beta * (1 + gy) * k[0] - uvar[0] * KY * (gy + 1 + beta * (delta - 1)) - k[-1] * KY * beta * (delta - 1) = 0
lvar[0] * phi_y * (rho_smooth - 1) - mpvar[0] + (rho_smooth - 1) * phi_m * pivar[0] + Rnomvar[0] - rho_smooth * Rnomvar[-1] = 0
rfvar[0] - Rnomvar[0] + pivar[1] = 0
chird[0] * sslabsr ^ rhoz * varphibar + varphibar * sslabsr ^ rhoz * rhoz * labsr[0] + (( - 1) - ga) * z[0] + (1 + ga) * z[-1] = 0
z[-1] + ctil[1] + chird[0] - ctil[0] + labsr[0] * (rhoz - 1) - ws[0] - z[0] + jz[1] = 0
( - ((1 + ga) * jz[0] * JZY * ssa)) - beta * a[0] * lam * (1 - obso) * ssz * VAY - ssa * (1 + ga) * labsaz[0] * sslabsaz * WSY - WSY * ws[0] * sslabsaz * (1 + ga) * ssa - ssa * ctil[0] * (1 + ga) * (JZY + sslabsaz * WSY) + ssa * lamvar[0] * (JZY * (1 + ga - beta * (1 - obso)) + WSY * (1 + ga) * sslabsaz) + z[0] * (VAY * ssz * (1 - obso) * beta * lam - ssa * (1 + ga) * JZY - WSY * sslabsaz * (1 + ga) * ssa) + ctil[1] * (1 + ga) * ssa * (JZY + sslabsaz * WSY) + jz[1] * (WSY * sslabsaz * (1 + ga) * ssa + ssa * (1 + ga) * JZY - VAY * ssz * (1 - obso) * beta * lam) + VAY * ssz * (1 - obso) * beta * lam * va[1] + z[-1] * (1 + ga) * ssa * (JZY + sslabsaz * WSY) = 0
labsaz[0] * rholambda - lamvar[0] = 0
ssa * rholambda * labsaz[0] * (ga + obso) - ssa * (1 + ga) * a[0] - a[-1] * ssa * (1 - obso) * (lam - 1) + z[-1] * ssz * lam * (1 - obso) = 0
mcvar[0] * (( - 1) - ga) * (muin - 1) + VAY * (1 - obso) * muvar * muin * beta * a[0] + VAY * (1 - obso) * muvar * muin * beta * ctil[0] + VAY * muvar * muin * (1 + ga) * va[0] + y[0] * (( - 1) - ga) * (muin - 1) - a[-1] * VAY * (1 - obso) * muvar * beta * muin - ctil[1] * VAY * (1 - obso) * muvar * beta * muin - va[1] * VAY * (1 - obso) * muvar * beta * muin = 0
( - (VAY * ssz * beta * a[0] * lam * (1 - obso) * rholambda)) - WSY * sslabsaz * ssa * ctil[0] * (1 + ga) + WSY * sslabsaz * ssa * (1 + ga) * labsaz[0] * (rholambda - 1) - WSY * ws[0] * sslabsaz * (1 + ga) * ssa + z[0] * (VAY * ssz * (1 - obso) * beta * lam * rholambda - WSY * sslabsaz * (1 + ga) * ssa) + ctil[1] * WSY * sslabsaz * (1 + ga) * ssa + jz[1] * (WSY * sslabsaz * (1 + ga) * ssa - VAY * ssz * (1 - obso) * beta * lam * rholambda) + va[1] * VAY * ssz * (1 - obso) * beta * lam * rholambda + z[-1] * WSY * sslabsaz * (1 + ga) * ssa = 0
c[0] * CY + g[0] * GY + invs[0] * IY + pinv[0] * IY - y[0] = 0
ssz * labs[0] * sslabs - ssz * sslabsr * labsr[0] + ssz * labsaz[0] * (sslabsr - sslabs) + a[-1] * ssa * sslabsaz - z[-1] * ssa * sslabsaz = 0
pivar[0] * (1 + beta * indw) / (1 + beta) + w[0] * (1 + (1 - epsw) * (1 - beta * epsw) / (epsw * (1 + beta)) / (1 + phi / (1 + 1 / muw))) - 1 / (1 + beta) * w[-1] - beta / (1 + beta) * w[1] + (1 - epsw) * (1 - beta * epsw) / (epsw * (1 + beta)) / (1 + phi / (1 + 1 / muw)) * (ctil[0] - lvar[0] * phi) - indw / (1 + beta) * pivar[-1] - pivar[1] * beta / (1 + beta) - muwvar[0] = 0
pivar[0] * (1 + beta * indw) / (1 + beta) + ws[0] * (1 + (1 - epsw) * (1 - beta * epsw) / (epsw * (1 + beta)) / (1 + phi / (1 + 1 / muw))) - 1 / (1 + beta) * ws[-1] - beta / (1 + beta) * ws[1] + (1 - epsw) * (1 - beta * epsw) / (epsw * (1 + beta)) / (1 + phi / (1 + 1 / muw)) * (ctil[0] - labs[0] * phi) - indw / (1 + beta) * pivar[-1] - pivar[1] * beta / (1 + beta) = 0
pivar[0] = pivar[-1] * indp / (1 + beta * indp) + pivar[1] * beta / (1 + beta * indp) + mcvar[0] * (1 - beta * epsp) * (1 - epsp) / ((1 + beta * indp) * epsp) + mu[0]
rd_exp[0] = labsr[0] + ws[0]
adopt_exp[0] = ws[0] + labsaz[0] + ssa / (ssz - ssa) * (z[-1] - a[-1])
tfp[0] - rhotfp * tfp[-1] - etfp[x] = 0
pinv[0] - rhopk * pinv[-1] - epinv[x] = 0
g[0] - rhog * g[-1] - eg[x] = 0
mub[0] - rhob * mub[-1] - eb[x] = 0
mu[0] - rhomu * mu[-1] - emu[x] = 0
muwvar[0] - rhow * muwvar[-1] - emuw[x] = 0
mpvar[0] - rhomp * mpvar[-1] - empvar[x] = 0
chird[0] - rhochi * chird[-1] - echi[x] = 0
Dy[0] = gy * 100 + y[0] - y[-1]
Dc[0] = gy * 100 + c[0] - c[-1]
Di[0] = gy * 100 + invs[0] - invs[-1] + pinv[0] - pinv[-1]
lvarobs[0] = lvar[0]
Dw[0] = w[0] + gy * 100 - w[-1]
Rnomobs[0] = Rnomvar[0] + 100 * ((1 + gy) * (1 - mubss) / beta - 1)
pinfobs[0] = pivar[0] + piconst
DrdNSF[0] = rd_exp[0] + 100 * gy * 4 - AUX_ENDO_LAG_39_3[-1]
AUX_ENDO_LAG_39_1[0] = rd_exp[-1]
AUX_ENDO_LAG_39_2[0] = AUX_ENDO_LAG_39_1[-1]
AUX_ENDO_LAG_39_3[0] = AUX_ENDO_LAG_39_2[-1]
end
@parameters model_translated_from_dynare begin
# liquidity premium in steady state 50 bps annually
mubss = 0.005/4
delta = 0.02
GY = 0.16
lam = 0.05
muin = 1.18
muvar = 1.1
obso = 0.02
rholambda = 0.925
thetta = 1.35
varphibar = 0.05
rconst = 0.5
piconst = 0.5
constelab = 0.013
## NOTE estimated parameters (latter)
std_eg = 2.5595
std_epinv = 1.2712
std_eb = 0.22225
std_emu = 0.093091
std_etfp = 0.48905
std_empvar = 0.096492
std_emuw = 0.28711
std_echi = 2.2119
rhotfp = 0.95092
rhopk = 0.8999
rhob = 0.92619
rhomp = 0.47126
rhomu = 0.40463
rhog = 0.9681
rhow = 0.27224
rhochi = 0.80216
rho_smooth = 0.82854
phi_m = 1.5745
phi_y = 0.38614
phi = 2.7026
ddac = 5.6386
epsp = 0.9346
epsw = 0.90568
indp = 0.24118
indw = 0.38198
h = 0.48303
edp = 4.0442
rhoz = 0.37076
alpha = 0.20068
gyest = 0.45317
betaest = 0.42122
muwest = 0.15
Rf_zlb = 100*(-1 + (1 + gyest/100)*(betaest/100+1))
## NOTE: Local-scope variables in dynare
gy = gyest/100
beta = (1 + betaest/100)^(-1)
ssa = 1
ssR = (1-mubss)*(1+gy)/beta
KY = alpha/(muin*muvar*(ssR-1+delta))
IY = KY*(delta+gy)
CY = 1 - GY - IY
ga = -1 + (1 + gy)^((1 - alpha)/(-1 + thetta))
ssz = ((1 + ga + (-1 + lam)*(1 - obso))*ssa)/(lam*(1 - obso))
sslabsr = ((ga + obso)/varphibar)^(1/rhoz)
VAY = ((1 + ga)*(-1 + muin))/(muin*muvar*(1 + ga - beta*(1 - obso)))
sslabsaz = ((-1 - ga + beta*(1 - obso))*rholambda*sslabsr^(1 - rhoz))/(beta*(-1 + rholambda)*varphibar)
WSY = -((beta^2*lam*(1 - obso)*(-1 + rholambda)*sslabsr^(-1 + rhoz)*ssz*varphibar*VAY)/((1 + ga)*(1 + ga + beta*(1 - obso)*(-1 + lam - lam*rholambda))*ssa))
JZY = -((beta*lam*(1 - obso)*(-1 + rholambda)*ssz*VAY)/((1 + ga + beta*(1 - obso)*(-1 + lam - lam*rholambda))*ssa))
sslabs = sslabsaz + sslabsr - (ssa*sslabsaz)/ssz
muw = muwest+1
kw = (1-epsw)*(1-beta*epsw)/(epsw*(1+beta))/(1+phi/(1+(1/muw)))
endThe difference in equations between what you posted and what I posted should explain the error. PS: Also with this model there is an extensive set of equations defining parameters (see last part in parameters block) which appears to lead to long run times for the derivatives. I will fix that in the next release. |
|
Thanks a lot! Using this function worked. I'll inspect the two scripts to spot the differences between them. I really appreciate your help! |
|
The function will catch the plain vanilla case: |
|
I meant importing the parameter values using MacroModelling after the dynare translation. Assume that I have a dictionary with the parameters as bellow: params = Dict{Symbol, Float64}()
params[:mubss] = 0.005/4
params[:delta] = 0.02
params[:GY] = 0.16
params[:lam] = 0.05
params[:muin] = 1.18
params[:muvar] = 1.1
params[:obso] = 0.02
params[:rholambda] = 0.925
params[:thetta] = 1.35
params[:varphibar] = 0.05
params[:rconst] = 0.5
params[:piconst] = 0.5
params[:constelab] = 0.013
params[:std_eg] = 2.5595
params[:std_epinv] = 1.2712
params[:std_eb] = 0.22225
params[:std_emu] = 0.093091
params[:std_etfp] = 0.48905
params[:std_empvar] = 0.096492
params[:std_emuw] = 0.28711
params[:std_echi] = 2.2119
params[:rhotfp] = 0.95092
params[:rhopk] = 0.8999
params[:rhob] = 0.92619
params[:rhomp] = 0.47126
params[:rhomu] = 0.40463
params[:rhog] = 0.9681
params[:rhow] = 0.27224
params[:rhochi] = 0.80216
params[:rho_smooth] = 0.82854
params[:phi_m] = 1.5745
params[:phi_y] = 0.38614
params[:phi] = 2.7026
params[:ddac] = 5.6386
params[:epsp] = 0.9346
params[:epsw] = 0.90568
params[:indp] = 0.24118
params[:indw] = 0.38198
params[:h] = 0.48303
params[:edp] = 4.0442
params[:rhoz] = 0.37076
params[:alpha] = 0.20068
params[:gyest] = 0.45317
params[:betaest] = 0.42122
params[:muwest] = 0.15
params[:Rf_zlb] = 100*(-1 + (1 + params[:gyest]/100)*(params[:betaest]/100+1))
params[:gy] = params[:gyest]/100
params[:beta] = (1 + params[:betaest]/100)^(-1)
params[:ssa] = 1
params[:ssR] = (1-params[:mubss])*(1+params[:gy])/params[:beta]
params[:KY] = params[:alpha]/(params[:muin]*params[:muvar]*(params[:ssR]-1+params[:delta]))
params[:IY] = params[:KY]*(params[:delta]+params[:gy])
params[:CY] = 1 - params[:GY] - params[:IY]
params[:ga] = -1 + (1 + params[:gy])^((1 - params[:alpha])/(-1 + params[:thetta]))
params[:ssz] = ((1 + params[:ga] + (-1 + params[:lam])*(1 - params[:obso]))*params[:ssa])/(params[:lam]*(1 - params[:obso]))
params[:sslabsr] = ((params[:ga] + params[:obso])/params[:varphibar])^(1/params[:rhoz])
params[:VAY] = ((1 + params[:ga])*(-1 + params[:muin]))/(params[:muin]*params[:muvar]*(1 + params[:ga] - params[:beta]*(1 - params[:obso])))
params[:sslabsaz] = ((-1 - params[:ga] + params[:beta]*(1 - params[:obso]))*params[:rholambda]*params[:sslabsr]^(1 - params[:rhoz]))/(params[:beta]*(-1 + params[:rholambda])*params[:varphibar])
params[:WSY] = -((params[:beta]^2*params[:lam]*(1 - params[:obso])*(-1 + params[:rholambda])*params[:sslabsr]^(-1 + params[:rhoz])*params[:ssz]*params[:varphibar]*params[:VAY])/((1 + params[:ga])*(1 + params[:ga] + params[:beta]*(1 - params[:obso])*(-1 + params[:lam] - params[:lam]*params[:rholambda]))*params[:ssa]))
params[:JZY] = -((params[:beta]*params[:lam]*(1 - params[:obso])*(-1 + params[:rholambda])*params[:ssz]*params[:VAY])/((1 + params[:ga] + params[:beta]*(1 - params[:obso])*(-1 + params[:lam] - params[:lam]*params[:rholambda]))*params[:ssa]))
params[:sslabs] = params[:sslabsaz] + params[:sslabsr] - (params[:ssa]*params[:sslabsaz])/params[:ssz]
params[:muw] = params[:muwest]+1
params[:kw] = (1-params[:epsw])*(1-params[:beta]*params[:epsw])/(params[:epsw]*(1+params[:beta]))/(1+params[:phi]/(1+(1/params[:muw])))
By using this, I avoid hardcoding the parameters by hand (which would make the evaluation faster). The next stage is to feed other models with a bayesian estimation using a dictionary (or other object type). In the matlab code, the authors simply populate the global scope with the actual values of the parameters and continue the simulation. I was thinking in doing a similar stuff. SInce this is a different topic, feel free to close the issue. If I came up with an idea, I'll write it down. Thanks! |
|
In the parameters block you have to define them using the name used in the model block and providing a value or relationship so that only using this information a steady state can be found. For estimation and also for simply Check out the tutorials for how to work with models and how to estimate them. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hi again,
I am trying to replicate a model in which I have a dynare script to depart from (I can paste it here if it helps). I notice that the lags/leads of some variables causes the error
MethodError: Cannotconvertan object of type Nothing to an object of type Int64. I tried in different julia version and the error still persists. I manage to replicate the error with only two equations, so I imagine that it became easier to debug:Code that is not working:
However, when I change
k[-1]tok[0]anda[-1]toa[0], the error no longer raises. This error message appears for some (not for every) equation that has mixed time leads/lags.Also, since I am trying to replicate a dynare script, do you have any suggestion on how to define those local-scope variables from dynare? (If I am not wrong, are those expressions starting with
#). So far, I simply treated them as parameters defined by other parameters.Thank you again for your help and sorry for opening another issue.
Best
The text was updated successfully, but these errors were encountered: