Debian Codegraph @

tests/forecast/fs2000_forecast.mod

raw
1 /*
2 * This file replicates the estimation of the cash in advance model (termed M1
3 * in the paper) described in Frank Schorfheide (2000): "Loss function-based
4 * evaluation of DSGE models", Journal of Applied Econometrics, 15(6), 645-670.
5 *
6 * The data are taken from the replication package at
7 * http://dx.doi.org/10.15456/jae.2022314.0708799949
8 *
9 * The prior distribution follows the one originally specified in Schorfheide's
10 * paper. Note that the elicited beta prior for rho in the paper
11 * implies an asymptote and corresponding prior mode at 0. It is generally
12 * recommended to avoid this extreme type of prior.
13 *
14 * Because the data are already logged and we use the loglinear option to conduct
15 * a full log-linearization, we need to use the logdata option.
16 *
17 * The equations are taken from J. Nason and T. Cogley (1994): "Testing the
18 * implications of long-run neutrality for monetary business cycle models",
19 * Journal of Applied Econometrics, 9, S37-S70, NC in the following.
20 * Note that there is an initial minus sign missing in equation (A1), p. S63.
21 *
22 * This implementation was originally written by Michel Juillard. Please note that the
23 * following copyright notice only applies to this Dynare implementation of the
24 * model.
25 */
26
27 /*
28 * Copyright © 2004-2025 Dynare Team
29 *
30 * This file is part of Dynare.
31 *
32 * Dynare is free software: you can redistribute it and/or modify
33 * it under the terms of the GNU General Public License as published by
34 * the Free Software Foundation, either version 3 of the License, or
35 * (at your option) any later version.
36 *
37 * Dynare is distributed in the hope that it will be useful,
38 * but WITHOUT ANY WARRANTY; without even the implied warranty of
39 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
40 * GNU General Public License for more details.
41 *
42 * You should have received a copy of the GNU General Public License
43 * along with Dynare. If not, see <https://www.gnu.org/licenses/>.
44 */
45
46 var m ${m}$ (long_name='money growth')
47 P ${P}$ (long_name='Price level')
48 c ${c}$ (long_name='consumption')
49 e ${e}$ (long_name='capital stock')
50 W ${W}$ (long_name='Wage rate')
51 R ${R}$ (long_name='interest rate')
52 k ${k}$ (long_name='capital stock')
53 d ${d}$ (long_name='dividends')
54 n ${n}$ (long_name='labor')
55 l ${l}$ (long_name='loans')
56 A
57 gy_obs ${\Delta \ln GDP}$ (long_name='detrended capital stock')
58 gp_obs ${\Delta \ln P}$ (long_name='detrended capital stock')
59 y ${y}$ (long_name='detrended output')
60 dA ${\Delta A}$ (long_name='TFP growth')
61 ;
62 varexo e_a ${\epsilon_A}$ (long_name='TFP shock')
63 e_m ${\epsilon_M}$ (long_name='Money growth shock')
64 ;
65
66 parameters alp ${\alpha}$ (long_name='capital share')
67 bet ${\beta}$ (long_name='discount factor')
68 gam ${\gamma}$ (long_name='long-run TFP growth')
69 logmst ${\log(m^*)}$ (long_name='long-run money growth')
70 rho ${\rho}$ (long_name='autocorrelation money growth')
71 phi ${\phi}$ (long_name='labor weight in consumption')
72 del ${\delta}$ (long_name='depreciation rate')
73 ;
74
75 % roughly picked values to allow simulating the model before estimation
76 alp = 0.33;
77 bet = 0.99;
78 gam = 0.003;
79 logmst = log(1.011);
80 rho = 0.7;
81 phi = 0.787;
82 del = 0.02;
83
84 model;
85 [name='NC before eq. (1), TFP growth equation']
86 dA = exp(gam+e_a);
87 [name='NC eq. (2), money growth rate']
88 log(m) = (1-rho)*logmst + rho*log(m(-1))+e_m;
89 [name='NC eq. (A1), Euler equation']
90 -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;
91 [name='NC below eq. (A1), firm borrowing constraint']
92 W = l/n;
93 [name='NC eq. (A2), intratemporal labour market condition']
94 -(phi/(1-phi))*(c*P/(1-n))+l/n = 0;
95 [name='NC below eq. (A2), credit market clearing']
96 R = P*(1-alp)*exp(-alp*(gam+e_a))*k(-1)^alp*n^(-alp)/W;
97 [name='NC eq. (A3), credit market optimality']
98 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;
99 [name='NC eq. (18), aggregate resource constraint']
100 c+k = exp(-alp*(gam+e_a))*k(-1)^alp*n^(1-alp)+(1-del)*exp(-(gam+e_a))*k(-1);
101 [name='NC eq. (19), money market condition']
102 P*c = m;
103 [name='NC eq. (20), credit market equilibrium condition']
104 m-1+d = l;
105 [name='Definition TFP shock']
106 e = exp(e_a);
107 [name='Implied by NC eq. (18), production function']
108 y = k(-1)^alp*n^(1-alp)*exp(-alp*(gam+e_a));
109 [name='Observation equation GDP growth']
110 gy_obs = dA*y/y(-1);
111 [name='Observation equation price level']
112 gp_obs = (P/P(-1))*m(-1)/dA;
113 A= rho*A(-1)+e_a;
114 end;
115
116 shocks;
117 var e_a; stderr 0.01;
118 var e_m; stderr 0.005;
119 end;
120
121 steady_state_model;
122 dA = exp(gam);
123 gst = 1/dA;
124 m = exp(logmst);
125 khst = ( (1-gst*bet*(1-del)) / (alp*gst^alp*bet) )^(1/(alp-1));
126 xist = ( ((khst*gst)^alp - (1-gst*(1-del))*khst)/m )^(-1);
127 nust = phi*m^2/( (1-alp)*(1-phi)*bet*gst^alp*khst^alp );
128 n = xist/(nust+xist);
129 P = xist + nust;
130 k = khst*n;
131
132 l = phi*m*n/( (1-phi)*(1-n) );
133 c = m/P;
134 d = l - m + 1;
135 y = k^alp*n^(1-alp)*gst^alp;
136 R = m/bet;
137 W = l/n;
138 ist = y-c;
139 q = 1 - d;
140
141 e = 1;
142
143 gp_obs = m/dA;
144 gy_obs = dA;
145 end;
146
147
148 stoch_simul(order=1,irf=0);
149
150 forecast(periods=500,conf_sig=0.95) A;
151 %% get theoretical forecast error variance for AR(1) process:
152 var_analytical=NaN(1+500,1);
153 var_analytical(1)=0; %zero variance in initial period
154 for ii=1:500
155 var_analytical(ii+1)=var_analytical(ii)+(rho^(ii-1))^2*0.01^2;
156 end
157 std_analytical=sqrt(var_analytical);
158
159 if norm(oo_.forecast.Mean.A-0,Inf)>1e-9
160 error('Mean forecast does not match theoretical moments')
161 end
162 if norm(oo_.forecast.HPDsup.A-norminv(0.975)*std_analytical(2:end),Inf)>1e-9
163 error('Forecast variance does not match theoretical moments')
164 end
165
166 varobs A;
167
168 shocks;
169 var e_a=0.01^2;
170 var A=0.01^2;
171 end;
172
173 forecast(periods=500,conf_sig=0.95) A;
174 std_analytical_ME=sqrt(var_analytical+0.01^2);
175 if norm(oo_.forecast.Mean.A-0,Inf)>1e-9
176 error('Mean forecast with ME does not match theoretical moments')
177 end
178 if norm(oo_.forecast.HPDsup.A-norminv(0.975)*std_analytical(2:end),Inf)>1e-9
179 error('Forecast variance with ME does not match theoretical moments')
180 end
181 if norm(oo_.forecast.HPDsup_ME.A-norminv(0.975)*std_analytical_ME(2:end),Inf)>1e-9
182 error('Forecast variance with ME does not match theoretical moments')
183 end
184
185
186 forecast(periods=500,conf_sig=0.95);
187 std_analytical_ME=sqrt(var_analytical+0.01^2);
188 if norm(oo_.forecast.Mean.A-0,Inf)>1e-9
189 error('Mean forecast with ME does not match theoretical moments')
190 end
191 if norm(oo_.forecast.HPDsup.A-norminv(0.975)*std_analytical(2:end),Inf)>1e-9
192 error('Forecast variance with ME does not match theoretical moments')
193 end
194 if norm(oo_.forecast.HPDsup_ME.A-norminv(0.975)*std_analytical_ME(2:end),Inf)>1e-9
195 error('Forecast variance with ME does not match theoretical moments')
196 end