stepan
/
dynare
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Document lik_init=5 and add unit test for lik_init

time-shift
Johannes Pfeifer 2015-10-18 18:55:33 +02:00
parent b01867c47e
commit a7ed27b889
5 changed files with 250 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
doc/dynare.texi
View File

@ -4898,13 +4898,19 @@ variables
@item 2
For nonstationary models: a wide prior is used with an initial matrix
of variance of the error of forecast diagonal with 10 on the diagonal
of variance of the error of forecast diagonal with 10 on the diagonal
(follows the suggestion of @cite{Harvey and Phillips(1979)})
@item 3
For nonstationary models: use a diffuse filter (use rather the @code{diffuse_filter} option)
@item 4
The filter is initialized with the fixed point of the Riccati equation
@item 5
Use i) option 2 for the non-stationary elements by setting their initial variance in the
forecast error matrix to 10 on the diagonal and all covariances to 0 and ii) option 1 for the stationary elements.
@end table
@noindent
@ -13065,6 +13071,11 @@ Hansen, Nikolaus and Stefan Kern (2004): ``Evaluating the CMA Evolution Strategy
on Multimodal Test Functions''. In: @i{Eighth International Conference on Parallel
Problem Solving from Nature PPSN VIII, Proceedings}, Berlin: Springer, 282--291
@item
Harvey, Andrew C. and Garry D.A. Phillips (1979): ``Maximum likelihood estimation of
regression models with autoregressive-moving average disturbances,''
@i{Biometrika}, 66(1), 49--58
@item
Ireland, Peter (2004): ``A Method for Taking Models to the Data,''
@i{Journal of Economic Dynamics and Control}, 28, 1205--26

2
matlab/dsge_likelihood.m
View File

@ -496,7 +496,7 @@ switch DynareOptions.lik_init
indx_unstable = find(sum(abs(V),2)>1e-5);
stable = find(sum(abs(V),2)<1e-5);
nunit = length(eigenv) - nstable;
Pstar = options_.Harvey_scale_factor*eye(np);
Pstar = DynareOptions.Harvey_scale_factor*eye(nunit);
if kalman_algo ~= 2
kalman_algo = 1;
end

2
tests/Makefile.am
View File

@ -164,6 +164,8 @@ MODFILES = \
ms-sbvar/test_ms_variances_repeated_runs.mod \
ms-dsge/test_ms_dsge.mod \
kalman/lyapunov/fs2000_lyap.mod \
kalman/fs2000_ns_lik_init.mod \
kalman/fs2000_lik_init.mod \
kalman_filter_smoother/gen_data.mod \
kalman_filter_smoother/algo1.mod \
kalman_filter_smoother/algo2.mod \

137
tests/kalman/fs2000_lik_init.mod Normal file
View File

@ -0,0 +1,137 @@
/*
* This file is based on the cash in advance model described
* Frank Schorfheide (2000): "Loss function-based evaluation of DSGE models",
* Journal of Applied Econometrics, 15(6), 645-670.
*
* The equations are taken from J. Nason and T. Cogley (1994): "Testing the
* implications of long-run neutrality for monetary business cycle models",
* Journal of Applied Econometrics, 9, S37-S70.
* Note that there is an initial minus sign missing in equation (A1), p. S63.
*
* This implementation was written by Michel Juillard. Please note that the
* following copyright notice only applies to this Dynare implementation of the
* model.
*/
/*
* Copyright (C) 2004-2013 Dynare Team
*
* This file is part of Dynare.
*
* Dynare is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* Dynare is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with Dynare. If not, see <http://www.gnu.org/licenses/>.
*/
var m P c e W R k d n l gy_obs gp_obs y dA;
varexo e_a e_m;
parameters alp bet gam mst rho psi del theta;
alp = 0.33;
bet = 0.99;
gam = 0.003;
mst = 1.011;
rho = 0.7;
psi = 0.787;
del = 0.02;
theta=0;
model;
dA = exp(gam+e_a);
log(m) = (1-rho)*log(mst) + rho*log(m(-1))+e_m;
-P/(c(+1)*P(+1)*m)+bet*P(+1)*(alp*exp(-alp*(gam+log(e(+1))))*k^(alp-1)*n(+1)^(1-alp)+(1-del)*exp(-(gam+log(e(+1)))))/(c(+2)*P(+2)*m(+1))=0;
W = l/n;
-(psi/(1-psi))*(c*P/(1-n))+l/n = 0;
R = P*(1-alp)*exp(-alp*(gam+e_a))*k(-1)^alp*n^(-alp)/W;
1/(c*P)-bet*P*(1-alp)*exp(-alp*(gam+e_a))*k(-1)^alp*n^(1-alp)/(m*l*c(+1)*P(+1)) = 0;
c+k = exp(-alp*(gam+e_a))*k(-1)^alp*n^(1-alp)+(1-del)*exp(-(gam+e_a))*k(-1);
P*c = m;
m-1+d = l;
e = exp(e_a);
y = k(-1)^alp*n^(1-alp)*exp(-alp*(gam+e_a));
gy_obs = dA*y/y(-1);
gp_obs = (P/P(-1))*m(-1)/dA;
end;
steady_state_model;
dA = exp(gam);
gst = 1/dA;
m = mst;
khst = ( (1-gst*bet*(1-del)) / (alp*gst^alp*bet) )^(1/(alp-1));
xist = ( ((khst*gst)^alp - (1-gst*(1-del))*khst)/mst )^(-1);
nust = psi*mst^2/( (1-alp)*(1-psi)*bet*gst^alp*khst^alp );
n = xist/(nust+xist);
P = xist + nust;
k = khst*n;
l = psi*mst*n/( (1-psi)*(1-n) );
c = mst/P;
d = l - mst + 1;
y = k^alp*n^(1-alp)*gst^alp;
R = mst/bet;
W = l/n;
ist = y-c;
q = 1 - d;
e = 1;
gp_obs = m/dA;
gy_obs = dA;
end;
varobs gp_obs gy_obs;
shocks;
var e_a; stderr 0.014;
var e_m; stderr 0.005;
corr gy_obs,gp_obs = 0.5;
end;
steady;
estimated_params;
alp, 0.356;
gam, 0.0085;
del, 0.01;
stderr e_a, 0.035449;
stderr e_m, 0.008862;
corr e_m, e_a, 0;
stderr gp_obs, 1;
stderr gy_obs, 1;
corr gp_obs, gy_obs,0;
end;
options_.TeX=1;
options_.debug=1;
%%default
options_.lik_init=1;
estimation(mode_compute=4,order=1,datafile='../fs2000/fsdat_simul',smoother,filter_decomposition,forecast = 8,filtered_vars,filter_step_ahead=[1,3],irf=20) m P c e W R k d y gy_obs;
fval_algo_0=oo_.likelihood_at_initial_parameters;
options_.lik_init=2;
estimation(mode_compute=4,order=1,datafile='../fs2000/fsdat_simul',smoother,filter_decomposition,forecast = 8,filtered_vars,filter_step_ahead=[1,3],irf=20) m P c e W R k d y gy_obs;
fval_algo_0=oo_.likelihood_at_initial_parameters;
options_.lik_init=3;
estimation(mode_compute=4,order=1,datafile='../fs2000/fsdat_simul',smoother,filter_decomposition,forecast = 8,filtered_vars,filter_step_ahead=[1,3],irf=20) m P c e W R k d y gy_obs;
fval_algo_0=oo_.likelihood_at_initial_parameters;
options_.lik_init=4;
estimation(mode_compute=4,order=1,datafile='../fs2000/fsdat_simul',smoother,filter_decomposition,forecast = 8,filtered_vars,filter_step_ahead=[1,3],irf=20) m P c e W R k d y gy_obs;
fval_algo_0=oo_.likelihood_at_initial_parameters;
options_.lik_init=5;
estimation(mode_compute=4,order=1,datafile='../fs2000/fsdat_simul',smoother,filter_decomposition,forecast = 8,filtered_vars,filter_step_ahead=[1,3],irf=20) m P c e W R k d y gy_obs;
fval_algo_0=oo_.likelihood_at_initial_parameters;

98
tests/kalman/fs2000_ns_lik_init.mod Normal file
View File

@ -0,0 +1,98 @@
// See fs2000.mod in the examples/ directory for details on the model
// This version estimates the model in level rather than in growth rates
var m P c e W R k d n l gy_obs gp_obs Y_obs P_obs y dA;
varexo e_a e_m;
parameters alp bet gam mst rho psi del;
alp = 0.33;
bet = 0.99;
gam = 0.003;
mst = 1.011;
rho = 0.7;
psi = 0.787;
del = 0.02;
model;
dA = exp(gam+e_a);
log(m) = (1-rho)*log(mst) + rho*log(m(-1))+e_m;
-P/(c(+1)*P(+1)*m)+bet*P(+1)*(alp*exp(-alp*(gam+log(e(+1))))*k^(alp-1)*n(+1)^(1-alp)+(1-del)*exp(-(gam+log(e(+1)))))/(c(+2)*P(+2)*m(+1))=0;
W = l/n;
-(psi/(1-psi))*(c*P/(1-n))+l/n = 0;
R = P*(1-alp)*exp(-alp*(gam+e_a))*k(-1)^alp*n^(-alp)/W;
1/(c*P)-bet*P*(1-alp)*exp(-alp*(gam+e_a))*k(-1)^alp*n^(1-alp)/(m*l*c(+1)*P(+1)) = 0;
c+k = exp(-alp*(gam+e_a))*k(-1)^alp*n^(1-alp)+(1-del)*exp(-(gam+e_a))*k(-1);
P*c = m;
m-1+d = l;
e = exp(e_a);
y = k(-1)^alp*n^(1-alp)*exp(-alp*(gam+e_a));
gy_obs = dA*y/y(-1);
gp_obs = (P/P(-1))*m(-1)/dA;
Y_obs/Y_obs(-1) = gy_obs;
P_obs/P_obs(-1) = gp_obs;
end;
steady_state_model;
dA = exp(gam);
gst = 1/dA;
m = mst;
khst = ( (1-gst*bet*(1-del)) / (alp*gst^alp*bet) )^(1/(alp-1));
xist = ( ((khst*gst)^alp - (1-gst*(1-del))*khst)/mst )^(-1);
nust = psi*mst^2/( (1-alp)*(1-psi)*bet*gst^alp*khst^alp );
n = xist/(nust+xist);
P = xist + nust;
k = khst*n;
l = psi*mst*n/( (1-psi)*(1-n) );
c = mst/P;
d = l - mst + 1;
y = k^alp*n^(1-alp)*gst^alp;
R = mst/bet;
W = l/n;
ist = y-c;
q = 1 - d;
e = 1;
gy_obs = exp(gam);
gp_obs = exp(-gam);
Y_obs=gy_obs;
P_obs=gp_obs;
end;
shocks;
var e_a; stderr 0.014;
var e_m; stderr 0.005;
end;
check(nocheck);
estimated_params;
alp, 0.356;
gam, 0.0085;
del, 0.01;
stderr e_a, 0.035449;
stderr e_m, 0.008862;
corr e_m, e_a, 0;
stderr P_obs, 1;
stderr Y_obs, 1;
corr P_obs, Y_obs,0;
end;
varobs P_obs Y_obs;
observation_trends;
P_obs (log(mst)-gam);
Y_obs (gam);
end;
%%default
options_.lik_init=2;
estimation(kalman_algo=1,mode_compute=4,order=1,datafile='../fs2000/fsdat_simul',smoother,filter_decomposition,forecast = 8,filtered_vars,filter_step_ahead=[1,3],irf=20) m P c e W R k d y gy_obs;
options_.lik_init=3;
estimation(kalman_algo=3,mode_compute=4,order=1,datafile='../fs2000/fsdat_simul',smoother,filter_decomposition,forecast = 8,filtered_vars,filter_step_ahead=[1,3],irf=20) m P c e W R k d y gy_obs;
options_.lik_init=5;
estimation(kalman_algo=1,mode_compute=4,order=1,datafile='../fs2000/fsdat_simul',smoother,filter_decomposition,forecast = 8,filtered_vars,filter_step_ahead=[1,3],irf=20) m P c e W R k d y gy_obs;