Hi Professor Pfeifer
I build a two-country dsge model with simple feedback rules. The steady state values are provided in the steady state model block, and the dynare code can run. After I comment out
the feedback rules and turn to ramsey optimal policy, dynare reports the steady state value contains NaN or inf. I have checked the variables and parameters,they are all defined and evaluated, so I don’t know which part of the model get wrong. By the way, does dynare have the debug function ?
James
paper.mod (9.99 KB)
find_ws.m (15.2 KB)
The unstable version says
[quote]POLICY AND TRANSITION FUNCTIONS
V de
Constant 1.005000 1.002494
bg(-1) -1.330337 0.533541
bgs(-1) -0.007358 -0.019916
fr(-1) 5.502187 0.766531
bc(-1) 0.571921 -0.038824
k(-1) -1.370886 0.112155
k_s(-1) -0.213025 0.019345
PI_H(-1) 1.345373 -0.611676
PI_F(-1) 0.515744 -0.373010
MULT_22(-1) -1.973379 0.300496
MULT_29(-1) 197.690235 -40.842518
c(-1) 0.890554 -0.348794
c_s(-1) 0.044540 0.123680
R(-1) 30.165820 -2.047788
R_s(-1) -0.976821 -0.573657
ph(-1) 0.000508 -2.490408
pfs(-1) -0.196266 -0.847491
m(-1) 0.520963 -0.038554
rer(-1) 0 -1.092264
taul(-1) 2.600815 -0.457738
tauls(-1) 1.035225 -0.232173
eps_Rs(-1) 0.122245 -0.342608
eps_H(-1) -6.827140 1.201562
eps_F(-1) -2.846870 0.638476
MULT_9(-1) -2429.018649 132.737927
MULT_10(-1) -2966.117463 201.433460
MULT_12(-1) 8871.107173 -1834.241503
MULT_13(-1) 5183.164060 -1073.463212
MULT_25(-1) -1030.877458 213.096110
eta_l 1.924603 -0.338726
eta_ls 0.476204 -0.106800
eta_Rs -0.244490 0.685216
eta_H -13.654280 2.403125
eta_F 5.693740 -1.276953
All endogenous are constant or non stationary, not displaying correlations and auto-correlations
STEADY-STATE RESULTS:
V 1.005
de 1.00249
c 0.77206
c_s 0.863804
i 0.523022
i_s 0.301329
R 1.00701
R_s 1.00451
b 53.3652
b_s 0.771631
bg 0.250512
bgs 0.771631
fr 0
bc 53.1147
w 2.15786
w_s 2.5785
l 0.4
l_s 0.3
tr -0.0644489
tr_s 0.155876
q 1
q_s 1
r 0.022004
r_s 0.022004
k 26.1511
k_s 15.0664
PI 1.005
PI_s 1.0025
PI_H 1.005
PI_F 1.0025
y_H 2.12928
y_F 0.971377
ph 0.788245
pfs 1.32725
phd 22.8814
phn 28.9549
pfd 13.9169
pfn 10.4724
m 21.3949
m_s 41.5568
rer 0.917813
mc_h 0.675616
mc_f 1.13763
g 0.38331
g_s 0.124127
y 1.67839
y_s 1.28926
gdp 1.67839
gdp_s 1.28926
ca 0
ca_s 0
tb 0
tb_s 0
taul 0.37
tauls 0.23
eps_v 1
eps_Rs 1
eps_H 1
eps_F 1[/quote]
Hi, professor Pfeifer
I have a new question. Before I turn to ramsey optimal policy, I run the code with simple rules, the model diagnostics does not report any colinear relationship. But as I turn to ramsey policy,the dynare reports the following messages: Warning: Matrix is close to singular or badly scaled. Results may be inaccurate. RCOND = 3.060307e-021. So I run the model diagnostics after it, the feedback message said that there are two collinear equtions. It this possible ? Should I worry about this problem ?
[quote=“jameszhow1985”]Hi, professor Pfeifer
I have a new question. Before I turn to ramsey optimal policy, I run the code with simple rules, the model diagnostics does not report any colinear relationship. But as I turn to ramsey policy,the dynare reports the following messages: Warning: Matrix is close to singular or badly scaled. Results may be inaccurate. RCOND = 3.060307e-021. So I run the model diagnostics after it, the feedback message said that there are two collinear equtions. It this possible ? Should I worry about this problem ?[/quote]
Can someone tell me how to fix the collinearity problem I mentioned above ? Any suggestion is appreciated.
Try to understand the solution you find. Maybe there is an economic reason why the computed Ramsey policy looks like this.