Dynare.jl (GitHub - DynareJulia/Dynare.jl: A Julia rewrite of Dynare: solving, simulating and estimating DSGE models. ), a Julia rewriting of Dynare,
The following features are implemented: steady state, local
Documentation is available at https://DynareJulia.github.io/Dynare.jl
Thank you for reporting any error, question or desired additional feature.
27 Likes
Nice package ! Considering the post about the β1.5 languageβ and the original Matlab version of dynare, I was wondering about your opinion on this Matlab to Julia reimplementation
4 Likes
mcreel
May 21, 2024, 10:35am
3
Thanks very much for this project, I am looking forward to using it.
With the current version (0.9.11), I can run the simple RBC estimation example from Estimation problem in Julia - Dynare Julia - Dynare Forum without errors, but I canβt find the estimation results anywhere. I looked inside the context structure, as well as in the subdirectories that are created. Is this expected with 0.9.11?
Dear Michael,
thank you for reporting the problem. There is obviously something broken in v0.9.11 with estimation. I will release a correction soon.
Note the error message at the end of the run:
ArgumentError("matrix contains Infs or NaNs")
Best
Michel
mcreel
May 21, 2024, 12:29pm
5
OK, thanks. I will try to test various things in the future.
Dear @mcreel ,
I just released Dynare v0.9.12 that should solve the problems with the estimation code.
] update
You may have to restart Julia.. mod file so that you see the first order approximation resultsDynareJulia estimation example used in https://forum.dynare.org/t/estimation-problem-in-julia/25412 and https://discourse.julialang.org/t/ann-dynare-jl-is-now-ready-for-testing/108177/4 Β· GitHub
1 Like
Hi Michel,
estimation: Dynare.EstimationResults([0.9208808863932328, 0.04441817638943685], [2.454376213054909, -3.114106515175857], [0.03408840700653537, 0.011260134441262335], [0.4220899442337999, 0.22912707304878288], [0.0011620194922432098 -0.00036399187020596423; -0.00036399187020596423 0.00012679062763530226], [0.17815992102329234 -0.09061689673115145; -0.09061689673115145 0.052499215603902284], 0)
which seems to be the estimation output. I donβt know quite how to interpret it, but the estimation seems to work. Am I supposed to see a nice table of estimated coefficients, standard devs., etc., as in Dynare on Octave? That is not appearing.
Next week, I will try MCMC, etcβ¦
Thanks!
That is very strange. I get the following output. The estimation results are at the end of the log:
julia> context = @dynare "RBCmj1";
Dynare version: 0.9.12
2024-05-25T19:22:41.091: Starting @dynare RBCmj1
["RBCmj1.mod", "language=julia", "json=compute"]
Dynare preprocessor version: 6.4.0+0
Starting preprocessing of the model file ...
Found 9 equation(s).
Evaluating expressions...
Computing static model derivatives (order 1).
Normalizing the static model...
Finding the optimal block decomposition of the static model...
3 block(s) found:
2 recursive block(s) and 1 simultaneous block(s).
the largest simultaneous block has 7 equation(s)
and 5 feedback variable(s).
Computing dynamic model derivatives (order 1).
Normalizing the dynamic model...
Finding the optimal block decomposition of the dynamic model...
3 block(s) found:
2 recursive block(s) and 1 simultaneous block(s).
the largest simultaneous block has 7 equation(s)
and 6 feedback variable(s).
JSON written after Computing step.
Preprocessing completed.
2024-05-25T19:22:41.121: End of preprocessing
2024-05-25T19:22:41.152: Start parse_statements!
Steady state
ββββββββ¬βββββββββ
Y β 1.21132
C β 0.9094
K β 12.0783
L β 0.3511
A β 1.0
R β 0.0351
W β 2.2423
I β 0.302
OBS_C β 0.0
ββββββββ΄βββββββββ
Coefficients of approximate solution function (reduced form)
ββββββββββ¬ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
β Y_t C_t K_t L_t A_t R_t W_t I_t OBS_C_t
ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
Ο(C) β -0.0 -0.0 0.0 0.0 0.0 -0.0 0.0 -0.0 -1.0997
Ο(K) β 0.0188 0.0431 0.9507 -0.0073 0.0 -0.0024 0.0812 -0.0243 0.0474
Ο(A) β 1.6994 0.2401 1.4593 0.2595 0.9226 0.0492 1.4888 1.4593 0.264
eps_A_t β 1.842 0.2602 1.5818 0.2813 1.0 0.0534 1.6137 1.5818 0.2862
ββββββββββ΄ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
Note: Ο(x) = x_{t-1} - steady_state(x)
THEORETICAL MOMENTS
βββββββββββ¬ββββββββββββββββββββββββββββββ
VARIABLE β MEAN STD. DEV. VARIANCE
βββββββββββββββββββββββββββββββββββββββββ
Y β 1.2113 0.2381 0.0567
C β 0.9094 0.1181 0.014
K β 12.0783 2.2739 5.1705
L β 0.3511 0.0259 0.0007
A β 1.0 0.1141 0.013
R β 0.0351 0.0052 0.0
W β 2.2423 0.3294 0.1085
I β 0.302 0.154 0.0237
OBS_C β 0.0 0.0136 0.0002
βββββββββββ΄ββββββββββββββββββββββββββββββ
VARIANCE DECOMPOSITION (in percent)
βββββββββββ¬βββββββ
VARIABLE β eps_A
ββββββββββββββββββ
Y β 1.0
C β 1.0
K β 1.0
L β 1.0
A β 1.0
R β 1.0
W β 1.0
I β 1.0
OBS_C β 1.0
βββββββββββ΄βββββββ
CORRELATION MATRIX
ββββββββ¬βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
β Y C K L A R W I OBS_C
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
Y β 1.0 0.8339 0.7491 0.7759 0.9895 0.4311 0.9485 0.906 0.4678
C β 0.8339 1.0 0.9903 0.2989 0.7455 -0.1385 0.9658 0.5219 0.0523
K β 0.7491 0.9903 1.0 0.1634 0.6457 -0.2748 0.9204 0.3983 -0.055
L β 0.7759 0.2989 0.1634 1.0 0.8589 0.9037 0.5362 0.97 0.7491
A β 0.9895 0.7455 0.6457 0.8589 1.0 0.5568 0.8929 0.9576 0.5513
R β 0.4311 -0.1385 -0.2748 0.9037 0.5568 1.0 0.123 0.7725 0.7539
W β 0.9485 0.9658 0.9204 0.5362 0.8929 0.123 1.0 0.7252 0.2499
I β 0.906 0.5219 0.3983 0.97 0.9576 0.7725 0.7252 1.0 0.6829
OBS_C β 0.4678 0.0523 -0.055 0.7491 0.5513 0.7539 0.2499 0.6829 1.0
ββββββββ΄βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
AUTOCORRELATION COEFFICIENTS
ββββββββ¬ββββββββββββββββββββββββββββββββββββββββ
β 1 2 3 4 5
ββββββββββββββββββββββββββββββββββββββββββββββββ
Y β 0.9403 0.8843 0.8318 0.7826 0.7365
C β 0.9945 0.9859 0.9746 0.961 0.9454
K β 0.998 0.9924 0.9837 0.9723 0.9586
L β 0.8773 0.7664 0.6662 0.5757 0.4942
A β 0.9226 0.8512 0.7853 0.7245 0.6684
R β 0.8867 0.7839 0.6907 0.6064 0.5301
W β 0.9764 0.952 0.9269 0.9014 0.8756
I β 0.8919 0.7937 0.7046 0.6237 0.5504
OBS_C β 0.2849 0.2469 0.2126 0.1817 0.154
ββββββββ΄ββββββββββββββββββββββββββββββββββββββββ
* Status: success
* Candidate solution
Final objective value: -2.243469e+02
* Found with
Algorithm: BFGS
* Convergence measures
|x - x'| = 2.33e-02 β° 0.0e+00
|x - x'|/|x'| = 7.47e-03 β° 0.0e+00
|f(x) - f(x')| = 1.80e-03 β° 0.0e+00
|f(x) - f(x')|/|f(x')| = 8.03e-06 β€ 1.0e-05
|g(x)| = 5.74e-01 β° 1.0e-08
* Work counters
Seconds run: 0 (vs limit Inf)
Iterations: 2
f(x) calls: 5
βf(x) calls: 5
Posterior mode
βββββββββββββ¬ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
Parameter β Estimated value Standard error 80% confidence interval
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
rhoA β 0.9209 0.0341 0.8772 .. 0.9645
Std(eps_A) β 0.0444 0.0113 0.03 .. 0.0588
βββββββββββββ΄ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
2024-05-25T19:22:50.589: End parse_statements!
2024-05-25T19:22:50.589: End parser
2 Likes
I simply forgot the semicolon at the end of
context = @dynare "RBCmj1";
With that, everything is fine. Many thanks!
Edit: running it on my own model, I can say that it is impressively fast compared to Dynare on Octave.
mcreel
May 26, 2024, 5:47pm
11
Do you prefer questions about Dynare.jl here, or on the Dynare forum? I do have a few questions.
Regarding testing, I have done MCMC, order=1, with a model of my own, no problems.
Both are OK for me, but for the users it would be better to have them in a single place. Letβs have the questions on the Dynare forum (https://forum.dynare.org ).
1 Like
