Page 1 of 1
Shocks and steady state convergence
Posted:
Wed Apr 30, 2014 6:51 am
by macroresearch123
Hi Johannes,
I added in another shock to my model in order to try and capture some dynamics that I think are important, but the steady state does not compute now; in fact, the shock must be negative for the steady state to exist / solve. I attached the file, but to save you time I was wondering if there are any diagnostics you suggest I run, i.e. commands that might reveal where the problem is. Neither shocks are correlated (which was causing a problem 2 months ago).
If there's anything quick that comes to mind, let me know and I will look into it!
Thank you!
Re: Shocks and steady state convergence
Posted:
Sun May 04, 2014 2:50 pm
by costa
I could run the code without any problem.
Re: Shocks and steady state convergence
Posted:
Sun May 04, 2014 7:16 pm
by macroresearch123
That's very odd - for me, Dynare is definitely returning that the steady state cannot be solved. (Even more peculiar is the fact that the residuals on the FOCs are all close to zero.) Costa, do you have any options on Dynare that are not on the default -- I am not even aware what these options might be? I really appreciate your downloading the file and checking, hopefully a solution in sight soon since indeed nothing should be wrong about the model I have specified.
Re: Shocks and steady state convergence
Posted:
Sun May 04, 2014 8:01 pm
by costa
I noticed I ran on DYnare 4.3.3, this is the output I received:
>> dynare benchmarkcb
Configuring Dynare ...
[mex] Generalized QZ.
[mex] Sylvester equation solution.
[mex] Kronecker products.
[mex] Sparse kronecker products.
[mex] Local state space iteration (second order).
[mex] Bytecode evaluation.
[mex] k-order perturbation solver.
[mex] k-order solution simulation.
[mex] Quasi Monte-Carlo sequence (Sobol).
[mex] Markov Switching SBVAR.
Starting Dynare (version 4.3.3).
Starting preprocessing of the model file ...
WARNING: APPDATA environment variable not found.
Found 17 equation(s).
Evaluating expressions...done
Computing static model derivatives:
- order 1
Computing dynamic model derivatives:
- order 1
- order 2
Processing outputs ...done
Preprocessing completed.
Starting MATLAB/Octave computing.
STEADY-STATE RESULTS:
C 0.0445366
H 0.054434
K 0.0546216
E 0.0890863
S 0.0445431
y 0.0499988
w 0.364922
r 0.155858
p 0.242702
A_k 1.01715
A_e 0.364551
A_s 1
A_p 1
z_k 0.017
z_e -1.00909
z_s 0
z_p 0
Residuals of the static equations:
Equation number 1 : 0
Equation number 2 : 0
Equation number 3 : -1.0264e-007
Equation number 4 : 0
Equation number 5 : 0
Equation number 6 : 0
Equation number 7 : 0
Equation number 8 : 0
Equation number 9 : 0
Equation number 10 : 0
Equation number 11 : 0
Equation number 12 : 0
Equation number 13 : 0
Equation number 14 : 0
Equation number 15 : 0
Equation number 16 : 0
Equation number 17 : 0
STEADY-STATE RESULTS:
C 0.0445366
H 0.054434
K 0.0546216
E 0.0890863
S 0.0445431
y 0.0499988
w 0.364922
r 0.155858
p 0.242702
A_k 1.01715
A_e 0.364551
A_s 1
A_p 1
z_k 0.017
z_e -1.00909
z_s 0
z_p 0
EIGENVALUES:
Modulus Real Imaginary
0.6587 0.6587 0
0.7 0.7 0
0.8 0.8 0
0.96 0.96 0
0.98 0.98 0
1.834 1.834 0
Inf Inf 0
Inf Inf 0
Inf -Inf 0
There are 4 eigenvalue(s) larger than 1 in modulus
for 4 forward-looking variable(s)
The rank condition is verified.
Total computing time : 0h00m24s
>> dynare benchmarkcb
Configuring Dynare ...
[mex] Generalized QZ.
[mex] Sylvester equation solution.
[mex] Kronecker products.
[mex] Sparse kronecker products.
[mex] Local state space iteration (second order).
[mex] Bytecode evaluation.
[mex] k-order perturbation solver.
[mex] k-order solution simulation.
[mex] Quasi Monte-Carlo sequence (Sobol).
[mex] Markov Switching SBVAR.
Starting Dynare (version 4.3.3).
Starting preprocessing of the model file ...
WARNING: APPDATA environment variable not found.
Found 17 equation(s).
Evaluating expressions...done
Computing static model derivatives:
- order 1
Computing dynamic model derivatives:
- order 1
- order 2
Processing outputs ...done
Preprocessing completed.
Starting MATLAB/Octave computing.
STEADY-STATE RESULTS:
C 0.0445366
H 0.054434
K 0.0546216
E 0.0890863
S 0.0445431
y 0.0499988
w 0.364922
r 0.155858
p 0.242702
A_k 1.01715
A_e 0.364551
A_s 1
A_p 1
z_k 0.017
z_e -1.00909
z_s 0
z_p 0
Residuals of the static equations:
Equation number 1 : 0
Equation number 2 : 0
Equation number 3 : -1.0264e-007
Equation number 4 : 0
Equation number 5 : 0
Equation number 6 : 0
Equation number 7 : 0
Equation number 8 : 0
Equation number 9 : 0
Equation number 10 : 0
Equation number 11 : 0
Equation number 12 : 0
Equation number 13 : 0
Equation number 14 : 0
Equation number 15 : 0
Equation number 16 : 0
Equation number 17 : 0
STEADY-STATE RESULTS:
C 0.0445366
H 0.054434
K 0.0546216
E 0.0890863
S 0.0445431
y 0.0499988
w 0.364922
r 0.155858
p 0.242702
A_k 1.01715
A_e 0.364551
A_s 1
A_p 1
z_k 0.017
z_e -1.00909
z_s 0
z_p 0
EIGENVALUES:
Modulus Real Imaginary
0.6587 0.6587 0
0.7 0.7 0
0.8 0.8 0
0.96 0.96 0
0.98 0.98 0
1.834 1.834 0
Inf Inf 0
Inf Inf 0
Inf -Inf 0
There are 4 eigenvalue(s) larger than 1 in modulus
for 4 forward-looking variable(s)
The rank condition is verified.
MODEL SUMMARY
Number of variables: 17
Number of stochastic shocks: 4
Number of state variables: 5
Number of jumpers: 4
Number of static variables: 8
Warning: Log of zero.
> In log10 at 20
In dyntable at 33
In stoch_simul at 105
In benchmarkcb at 303
In dynare at 120
MATRIX OF COVARIANCE OF EXOGENOUS SHOCKS
Variables eps_k eps_e eps_s eps_p
eps_k 0.000900 0.000000 0.000000 0.000000
eps_e 0.000000 0.000900 0.000000 0.000000
eps_s 0.000000 0.000000 0.000900 0.000000
eps_p 0.000000 0.000000 0.000000 0.000900
POLICY AND TRANSITION FUNCTIONS
C H K E S y w r p A_k A_e A_s A_p z_k z_e z_s z_p
Constant 0.044521 0.054464 0.054657 0.089120 0.044560 0.050018 0.364862 0.155918 0.242702 1.017145 0.364551 1.000000 1.000000 0.017000 -1.009090 0 0
(correction) -0.000016 0.000030 0.000035 0.000034 0.000017 0.000019 -0.000060 0.000060 0 0 0 0 0 0 0 0 0
K(-1) 0.384556 -0.175335 0.658739 0.211866 0.105933 0.143295 1.543210 -2.696337 0.710267 0 0 0 0 0 0 0 0
z_k(-1) 0.040925 0.020190 0.017740 0.096774 0.048387 0.058665 0.171830 0.131198 0.126733 0.996802 0 0 0 0.980000 0 0 0
z_e(-1) 0.013224 0.047347 0.020839 -0.016364 -0.008182 0.034063 0.063335 0.162615 0.094591 0 0.349969 0 0 0 0.960000 0 0
z_s(-1) -0.000015 -0.000012 0.000007 -0.000014 0.031173 -0.000008 0.000025 -0.000024 0 0 0 0.700000 0 0 0 0.700000 0
z_p(-1) 0.009511 0.036046 0.016372 0.052446 0.026223 0.025883 0.046025 0.122867 -0.103456 0 0 0 0.800000 0 0 0 0.800000
eps_k 0.041760 0.020602 0.018102 0.098749 0.049375 0.059862 0.175337 0.133875 0.129320 1.017145 0 0 0 1.000000 0 0 0
eps_e 0.013775 0.049319 0.021707 -0.017045 -0.008523 0.035482 0.065974 0.169391 0.098532 0 0.364551 0 0 0 1.000000 0 0
eps_s -0.000021 -0.000017 0.000010 -0.000020 0.044533 -0.000011 0.000035 -0.000035 0 0 0 1.000000 0 0 0 1.000000 0
eps_p 0.011888 0.045058 0.020465 0.065558 0.032779 0.032354 0.057531 0.153583 -0.129320 0 0 0 1.000000 0 0 0 1.000000
K(-1),K(-1) -1.336293 0.376306 -0.767210 -3.430573 -1.715287 -2.103502 -8.943975 44.498413 -6.011024 0 0 0 0 0 0 0 0
z_k(-1),K(-1) 0.220448 0.005525 -0.016075 0.306701 0.153351 0.204373 0.185178 -2.298949 0.175105 0 0 0 0 0 0 0 0
z_k(-1),z_k(-1) 0.019044 -0.013328 0.003582 0.032623 0.016311 0.022626 0.039142 0.005777 0.015622 0.488433 0 0 0 0 0 0 0
z_e(-1),K(-1) 0.106736 0.049012 0.129203 0.191606 0.095803 0.235939 0.245620 -2.245723 0.537140 0 0 0 0 0 0 0 0
z_e(-1),z_k(-1) 0.022151 0.028313 0.026289 -0.005519 -0.002760 0.048440 0.076551 0.187627 0.095842 0 0 0 0 0 0 0 0
z_e(-1),z_e(-1) -0.001606 0.006792 0.003404 -0.014606 -0.007303 0.001797 -0.007623 0.039719 -0.003833 0 0.167985 0 0 0 0 0 0
z_s(-1),K(-1) -0.000333 -0.000039 0.000260 -0.000123 0.074092 -0.000073 0.000262 0.000266 0 0 0 0 0 0 0 0 0
z_s(-1),z_k(-1) -0.000039 -0.000026 0.000016 -0.000040 0.033851 -0.000023 0.000055 -0.000064 0 0 0 0 0 0 0 0 0
z_s(-1),z_e(-1) -0.000006 -0.000018 -0.000004 -0.000006 -0.005730 -0.000010 0.000019 -0.000040 0 0 0 0 0 0 0 0 0
z_s(-1),z_s(-1) -0.000005 0.000001 0.000006 0.000001 0.010909 0 -0.000002 0.000002 0 0 0 0.245000 0 0 0 0 0
z_p(-1),K(-1) 0.071008 0.062618 0.125534 0.332973 0.166487 0.196542 0.161362 -1.628062 -0.142943 0 0 0 0 0 0 0 0
z_p(-1),z_k(-1) 0.016855 0.023343 0.021255 0.069021 0.034511 0.038110 0.059814 0.148272 -0.025505 0 0 0 0 0 0 0 0
z_p(-1),z_e(-1) -0.001388 0.010975 0.004645 -0.032930 -0.016465 0.003257 -0.006826 0.063518 -0.078238 0 0 0 0 0 0 0 0
z_p(-1),z_s(-1) -0.000015 -0.000011 0.000008 -0.000012 0.018350 -0.000006 0.000010 -0.000026 0 0 0 0 0 0 0 0 0
z_p(-1),z_p(-1) -0.000656 0.003152 0.001167 0.005122 0.002561 0.000510 -0.003189 0.020681 0.010410 0 0 0 0.320000 0 0 0 0
eps_k,eps_k 0.019829 -0.013877 0.003729 0.033968 0.016984 0.023559 0.040756 0.006016 0.016266 0.508573 0 0 0 0 0 0 0
eps_e,eps_k 0.023545 0.030094 0.027944 -0.005867 -0.002933 0.051489 0.081368 0.199434 0.101873 0 0 0 0 0 0 0 0
eps_e,eps_e -0.001743 0.007370 0.003693 -0.015849 -0.007924 0.001950 -0.008271 0.043098 -0.004159 0 0.182275 0 0 0 0 0 0
eps_s,eps_k -0.000057 -0.000038 0.000024 -0.000058 0.049345 -0.000034 0.000080 -0.000093 0 0 0 0 0 0 0 0 0
eps_s,eps_e -0.000008 -0.000026 -0.000006 -0.000008 -0.008527 -0.000015 0.000028 -0.000059 0 0 0 0 0 0 0 0 0
eps_s,eps_s -0.000011 0.000002 0.000012 0.000003 0.022263 0.000002 -0.000005 0.000005 0 0 0 0.500000 0 0 0 0 0
eps_p,eps_k 0.021499 0.029774 0.027112 0.088037 0.044019 0.048610 0.076294 0.189122 -0.032532 0 0 0 0 0 0 0 0
eps_p,eps_e -0.001807 0.014290 0.006048 -0.042877 -0.021439 0.004240 -0.008887 0.082706 -0.101873 0 0 0 0 0 0 0 0
eps_p,eps_s -0.000026 -0.000020 0.000015 -0.000021 0.032768 -0.000011 0.000017 -0.000046 0 0 0 0 0 0 0 0 0
eps_p,eps_p -0.001026 0.004926 0.001823 0.008003 0.004002 0.000797 -0.004983 0.032314 0.016266 0 0 0 0.500000 0 0 0 0
K(-1),eps_k 0.224947 0.005638 -0.016403 0.312961 0.156480 0.208544 0.188957 -2.345867 0.178679 0 0 0 0 0 0 0 0
K(-1),eps_e 0.111184 0.051054 0.134586 0.199590 0.099795 0.245770 0.255855 -2.339295 0.559520 0 0 0 0 0 0 0 0
K(-1),eps_s -0.000475 -0.000056 0.000371 -0.000176 0.105845 -0.000104 0.000374 0.000380 0 0 0 0 0 0 0 0 0
K(-1),eps_p 0.088760 0.078273 0.156918 0.416216 0.208108 0.245677 0.201703 -2.035077 -0.178679 0 0 0 0 0 0 0 0
z_k(-1),eps_k 0.038865 -0.027200 0.007310 0.066577 0.033289 0.046175 0.079882 0.011791 0.031882 0.996802 0 0 0 0 0 0 0
z_k(-1),eps_e 0.023074 0.029492 0.027385 -0.005749 -0.002875 0.050459 0.079741 0.195445 0.099836 0 0 0 0 0 0 0 0
z_k(-1),eps_s -0.000056 -0.000037 0.000023 -0.000057 0.048359 -0.000033 0.000078 -0.000091 0 0 0 0 0 0 0 0 0
z_k(-1),eps_p 0.021069 0.029179 0.026569 0.086276 0.043138 0.047638 0.074768 0.185340 -0.031882 0 0 0 0 0 0 0 0
z_e(-1),eps_k 0.022603 0.028890 0.026826 -0.005632 -0.002816 0.049429 0.078114 0.191457 0.097798 0 0 0 0 0 0 0 0
z_e(-1),eps_e -0.003346 0.014150 0.007091 -0.030430 -0.015215 0.003745 -0.015881 0.082748 -0.007986 0 0.349969 0 0 0 0 0 0
z_e(-1),eps_s -0.000008 -0.000025 -0.000006 -0.000008 -0.008186 -0.000014 0.000027 -0.000057 0 0 0 0 0 0 0 0 0
z_e(-1),eps_p -0.001735 0.013718 0.005806 -0.041162 -0.020581 0.004071 -0.008532 0.079398 -0.097798 0 0 0 0 0 0 0 0
z_s(-1),eps_k -0.000040 -0.000027 0.000017 -0.000041 0.034542 -0.000024 0.000056 -0.000065 0 0 0 0 0 0 0 0 0
z_s(-1),eps_e -0.000006 -0.000019 -0.000004 -0.000006 -0.005969 -0.000010 0.000019 -0.000042 0 0 0 0 0 0 0 0 0
z_s(-1),eps_s -0.000015 0.000003 0.000017 0.000004 0.031168 0.000002 -0.000007 0.000007 0 0 0 0.700000 0 0 0 0 0
z_s(-1),eps_p -0.000018 -0.000014 0.000011 -0.000015 0.022938 -0.000008 0.000012 -0.000032 0 0 0 0 0 0 0 0 0
z_p(-1),eps_k 0.017199 0.023819 0.021689 0.070430 0.035215 0.038888 0.061035 0.151298 -0.026026 0 0 0 0 0 0 0 0
z_p(-1),eps_e -0.001446 0.011432 0.004838 -0.034302 -0.017151 0.003392 -0.007110 0.066165 -0.081498 0 0 0 0 0 0 0 0
z_p(-1),eps_s -0.000021 -0.000016 0.000012 -0.000017 0.026215 -0.000009 0.000014 -0.000037 0 0 0 0 0 0 0 0 0
z_p(-1),eps_p -0.001641 0.007881 0.002916 0.012805 0.006403 0.001275 -0.007974 0.051702 0.026026 0 0 0 0.800000 0 0 0 0
MOMENTS OF SIMULATED VARIABLES
VARIABLE MEAN STD. DEV. VARIANCE SKEWNESS KURTOSIS
C 0.043244 0.007469 0.000056 0.785069 0.471846
H 0.052933 0.004760 0.000023 0.484298 0.365033
K 0.052497 0.007859 0.000062 0.274655 -0.600424
E 0.089179 0.014859 0.000221 1.356122 2.783020
S 0.044621 0.007653 0.000059 1.107810 1.720243
y 0.048466 0.008371 0.000070 0.699098 0.284364
w 0.358068 0.030911 0.000955 0.538429 -0.038574
r 0.155421 0.010215 0.000104 0.228332 0.086434
p 0.236318 0.021202 0.000450 0.397393 -0.518065
A_k 1.017496 0.133240 0.017753 1.603384 3.758050
A_e 0.354252 0.041951 0.001760 1.093574 2.041339
A_s 1.000438 0.040012 0.001601 0.148727 -0.231260
A_p 1.003968 0.052358 0.002741 0.340672 0.036812
z_k 0.009773 0.122841 0.015090 1.120865 1.937258
z_e -1.044526 0.114523 0.013116 0.625939 1.098309
z_s -0.000361 0.039977 0.001598 0.044158 -0.251859
z_p 0.002616 0.051926 0.002696 0.195138 -0.087193
CORRELATION OF SIMULATED VARIABLES
VARIABLE C H K E S y w r p A_k A_e A_s A_p z_k z_e z_s z_p
C 1.0000 0.4833 0.8754 0.8401 0.8194 0.9900 0.9978 0.0365 0.8756 0.8411 0.2331 0.0447 0.2197 0.8400 0.2381 0.0436 0.2210
H 0.4833 1.0000 0.7846 0.0579 0.0705 0.5632 0.5240 0.4434 0.4967 0.0039 0.8589 0.0520 0.4289 0.0111 0.8619 0.0518 0.4264
K 0.8754 0.7846 1.0000 0.5079 0.5002 0.8816 0.8951 0.0168 0.8151 0.4771 0.6169 0.0471 0.3209 0.4809 0.6195 0.0470 0.3196
E 0.8401 0.0579 0.5079 1.0000 0.9712 0.8202 0.8193 0.0633 0.5674 0.9654 -0.3044 0.0345 0.2964 0.9637 -0.3022 0.0328 0.2999
S 0.8194 0.0705 0.5002 0.9712 1.0000 0.8027 0.8016 0.0790 0.5597 0.9370 -0.2894 0.2685 0.2893 0.9392 -0.2862 0.2666 0.2931
y 0.9900 0.5632 0.8816 0.8202 0.8027 1.0000 0.9915 0.1706 0.8669 0.8158 0.2859 0.0520 0.2672 0.8157 0.2910 0.0507 0.2685
w 0.9978 0.5240 0.8951 0.8193 0.8016 0.9915 1.0000 0.0535 0.8804 0.8169 0.2712 0.0524 0.2357 0.8195 0.2768 0.0513 0.2370
r 0.0365 0.4434 0.0168 0.0633 0.0790 0.1706 0.0535 1.0000 0.0184 0.0455 0.1935 0.0526 0.2849 0.0539 0.1995 0.0504 0.2867
p 0.8756 0.4967 0.8151 0.5674 0.5597 0.8669 0.8804 0.0184 1.0000 0.6766 0.4526 0.0540 -0.2022 0.6821 0.4624 0.0532 -0.2019
A_k 0.8411 0.0039 0.4771 0.9654 0.9370 0.8158 0.8169 0.0455 0.6766 1.0000 -0.2663 0.0318 0.0555 0.9955 -0.2646 0.0300 0.0595
A_e 0.2331 0.8589 0.6169 -0.3044 -0.2894 0.2859 0.2712 0.1935 0.4526 -0.2663 1.0000 0.0154 0.0206 -0.2714 0.9956 0.0161 0.0168
A_s 0.0447 0.0520 0.0471 0.0345 0.2685 0.0520 0.0524 0.0526 0.0540 0.0318 0.0154 1.0000 0.0114 0.0447 0.0194 0.9997 0.0137
A_p 0.2197 0.4289 0.3209 0.2964 0.2893 0.2672 0.2357 0.2849 -0.2022 0.0555 0.0206 0.0114 1.0000 0.0548 0.0126 0.0111 0.9994
z_k 0.8400 0.0111 0.4809 0.9637 0.9392 0.8157 0.8195 0.0539 0.6821 0.9955 -0.2714 0.0447 0.0548 1.0000 -0.2682 0.0429 0.0588
z_e 0.2381 0.8619 0.6195 -0.3022 -0.2862 0.2910 0.2768 0.1995 0.4624 -0.2646 0.9956 0.0194 0.0126 -0.2682 1.0000 0.0198 0.0086
z_s 0.0436 0.0518 0.0470 0.0328 0.2666 0.0507 0.0513 0.0504 0.0532 0.0300 0.0161 0.9997 0.0111 0.0429 0.0198 1.0000 0.0134
z_p 0.2210 0.4264 0.3196 0.2999 0.2931 0.2685 0.2370 0.2867 -0.2019 0.0595 0.0168 0.0137 0.9994 0.0588 0.0086 0.0134 1.0000
AUTOCORRELATION OF SIMULATED VARIABLES
VARIABLE 1 2 3 4 5
C 0.9797 0.9518 0.9185 0.8834 0.8456
H 0.8918 0.8137 0.7506 0.7045 0.6718
K 0.9850 0.9555 0.9182 0.8779 0.8369
E 0.9652 0.9314 0.8950 0.8617 0.8265
S 0.9475 0.8998 0.8505 0.8030 0.7579
y 0.9581 0.9168 0.8750 0.8368 0.7970
w 0.9783 0.9487 0.9138 0.8777 0.8396
r 0.5752 0.3120 0.1379 0.0513 0.0157
p 0.9523 0.9056 0.8617 0.8183 0.7771
A_k 0.9709 0.9433 0.9145 0.8871 0.8569
A_e 0.9672 0.9389 0.9131 0.8849 0.8566
A_s 0.6875 0.4561 0.2807 0.1678 0.0759
A_p 0.7991 0.6334 0.4838 0.3701 0.2913
z_k 0.9682 0.9371 0.9051 0.8756 0.8451
z_e 0.9643 0.9342 0.9075 0.8779 0.8489
z_s 0.6883 0.4590 0.2845 0.1718 0.0796
z_p 0.7999 0.6354 0.4876 0.3744 0.2966
Total computing time : 0h00m37s
When I tried to run on dynare 4.2.2 I got an error message, as you mentioned. So I believe the problem is with the new version.
Re: Shocks and steady state convergence
Posted:
Sun May 04, 2014 8:11 pm
by costa
I tried with different solve_algo configurations and the error remains. Why donĀ“t you try to alter the initval block with the values oobtained with dynare 4.3?
Re: Shocks and steady state convergence
Posted:
Mon May 05, 2014 2:44 am
by macroresearch123
Indeed, I run it all on 4.3.3 as well. I also tried all the solve_algo options (0,1,2,3). I wonder if there are any Dynare specialists who were involved in the project still checking the forum?
EDIT: I meant to say that I tried running it on 4.3.3 (my old computer) too. Will keep checking though...
Re: Shocks and steady state convergence
Posted:
Wed May 07, 2014 9:18 am
by jpfeifer
Could you give me an update what the current problem is?
Re: Shocks and steady state convergence
Posted:
Wed May 07, 2014 2:47 pm
by macroresearch123
That's precisely it -- I cannot find anything wrong with the code, but when I compute the model Dynare says that the steady state does not exist / is impossible to find. (I attached the file again, but it's almost the same as the first version I attached in the post in this thread.) However, Costa was able to get the model to run, and I coded a simple Matlab file to check for the steady state using Fsolve (and it exists).
Re: Shocks and steady state convergence
Posted:
Wed May 07, 2014 3:31 pm
by jpfeifer
Did you provide the steady state found with fsolve in initval?
Re: Shocks and steady state convergence
Posted:
Wed May 07, 2014 5:55 pm
by macroresearch123
Yes, I used the initial conditions that I obtained from the Fsolve -- I didn't link the two files, I just copied them over. (FYI when I used Fsolve, the "trust region dogleg" was the one that worked, but levenberg-marquardt and trust region reflection did not work; in fact, levenberg-marquardt said that there was a unit root with the "exitflag" equal to -1 result.)
If there's a suggestion on linking the Fsolve file with the Dynare code, I'd definitely be interested. My concern is that Dynare does so much, but on the user's part it's hard to know what's actually going on!
EDIT: I thought of re-downloading 4.3.3 and trying it instead of my 4.4.2. Which one of the version 3's on the site is the right to download?
http://www.dynare.org/release/oldies/dy ... b/windows/ I looked through, but I wasn't sure what each of these differentiated version 3 names were about.
Re: Shocks and steady state convergence
Posted:
Thu May 08, 2014 8:22 am
by jpfeifer
I am still not following. The initial values you provide are definitely not a steady state. Put resid(1) before steady to see that there are quite big residuals. I am not sure what you did solve with fsolve.
Regarding 4.3.3, it's at
http://www.dynare.org/release/windows/dynare-4.3.3-win.exe
Re: Shocks and steady state convergence
Posted:
Thu May 08, 2014 2:18 pm
by macroresearch123
If I use Fsolve and set the initial values equal to what I am told, I get the following residuals
--resids
Equation number 1 : 0
Equation number 2 : 0.064761
Equation number 3 : 0.0012457
Equation number 4 : 0.29732
Equation number 5 : 0.016403
Equation number 6 : 0.016832
Equation number 7 : -0.036375
Equation number 8 : 0.23757
Equation number 9 : 0
Equation number 10 : 0
Equation number 11 : 0
Equation number 12 : 0
Equation number 13 : 0
Equation number 14 : 0
Equation number 15 : 0
Equation number 16 : 0
Equation number 17 : 0
--
These residuals are small. If I use the prior SS initial values, equal to below, then I get even smaller residuals on the scale of .00006. The fundamental problem is that another user was able to run the code, however, I was not -- and we are talking about the same exact code.
p=.025; r=.15; w=.9;
C=0.77; H=0.3; K=.8; E=.7; S=0.3; y=1;
A_k=1; A_e=1; A_p=1; A_s=1;
eps_k=0; eps_e=0; eps_s=0; eps_p=0;
EDIT: I used the 4.3.3 version that you wrote above (thank you) and the code worked, just as Costa had; this is peculiar since I tried the 4.3.3 on another computer and it did not work. Either way, if possible I would like to learn what the source of the problem could be since it seems I spend hours and hours searching for a problem in Dynare when really the problem is not one at all, such as that time with the correlated shocks or this.
Similarly, I also used the MH estimation to see how its results compare with estimation of my parameters that I did independently using GMM. However, it came up with an error -- googled it and appears to be a "known bug", but wasn't sure if there are tips on what is causing it, ie too many estimated vars, etc?
You did not declare endogenous variables after the estimation/calib_smoother command.
??? Error using ==> union
Too many input arguments.
Re: Shocks and steady state convergence
Posted:
Fri May 09, 2014 7:24 am
by jpfeifer
First of all, residual of 0.29 and 0.23 are not small. This indicates that the starting values are not good enough. You might be lucky that the solver you use finds the solution, but it might also simply get stuck at a local minimum. It seems that is what happens. The problem across versions and machines might be that they use different random numbers and many solvers randomly perturb the search direction when encountering a cliff. This might explain the observed behavior.
If the problem you get is a known bug
http://www.dynare.org/DynareWiki/KnownBugs, use the most recent version of Dynare for estimation. Simply use the steady state found in 4.3.3 as initial values for 4.4.2.
Re: Shocks and steady state convergence
Posted:
Sun May 11, 2014 7:22 pm
by rox17
`What if all the residuals are 0 and the steady state for the observed var the same ?
Re: Shocks and steady state convergence
Posted:
Mon May 12, 2014 9:02 am
by donihue
Macroresearch123:
Just to confirm that when I ran your model using Dynare 4.4.2 and steady(solve_algo=4,maxit=10000), it worked with the same results as obtained by costa.
Regards
Donihue