Dear Friends,
This year the Dynare Summer School will be hosted, from May 22 to May
26, 2023, by École Normale Supérieure (Campus Jourdan). The courses will
focus on simulation and estimation of DSGE models with Dynare, and Marco
Del Negro (Federal Reserve Bank of New York) will be the invited
speaker.
More information is available on our website:
https://www.dynare.org/events/dynare-summer-school-2023-applications-open
Applications should be sent to school(a)dynare.org (with a CV and a recent
research paper attached as PDFs) no later than March 3, 2023.
Best regards,
Stéphane.
Dear Dynare friends,
We are pleased to announce the release of Dynare 5.3.
This maintenance release fixes various bugs.
The Windows, macOS and source packages are already available for
download atthe Dynare website.
All users are strongly encouraged to upgrade.
This release is compatible with MATLAB versions ranging from 8.3
(R2014a) to 9.13 (R2022b), and with GNU Octave version 7.3.0 (under
Windows).
Note for macOS users with an Apple Silicon processor, and who are also
MATLAB users: the official MATLAB version for use on those processors
is still the Intel version (running through Rosetta 2), so the official
Dynare package for download on our website is built for Intel only.
However, since Mathworks has released a beta version of MATLAB for
Apple Silicon, we created a beta package of Dynare that you can try
with it. See this forum thread for more
details:https://forum.dynare.org/t/testers-wanted-release-of-dynare-5-x-bet…
Here is a list of the problems identified in version 5.2 and that have
been fixed in version 5.3:
* Problems with the steady_state operator:
if a steady_state operator contained an algebraic expression
appearing
multiple times in the model and sufficiently complex to trigger the
creation of a temporary term, then the result of the operator would
be
wrong (the operator was essentially ignored)if a steady_state
operator contained a call to an external function, then
the result of the operator would be wrong (the operator was
essentially
ignored). A proper fix to this problem would require substantial
architectural changes, so for now it is forbidden to use an external
function inside a steady_state operator
* Pruning in particle filtering at order 2 was not using the exact
same formula
as the original Kim et al. (2008) paper. A second-order term entered
the
cross-product between states and shocks, where it should have been a
first-order term. This however would not lead to explosive
trajectories in
practice
* The simul_replic option of the stoch_simul command would not store
the
binary file in the Output folder
* Problems with Ramsey policy (ramsey_model/ramsey_policy commands):
steady state files would not work when auxiliary variables included
Lagrange multipliersfor linear competitive equilibrium laws of
motion, welfare evaluated at
higher order was erroneously equated to steady state welfare
* The discretionary_policy command would not always correctly infer
the
number of instruments and equations, leading to spurious error
messages
* Perfect foresight simulations of purely forward or backward models
would
crash if complex numbers were encountered
* When using both block and bytecode options of the model block, if
the
model was such that a sufficiently complex algebraic expression
appeared both
in the residuals and in the derivatives, leading to the creation of
a
temporary term, then the results could be incorrect under some
circumstances
* When using the bytecode option of the model block, leads of more
than
+127 or lags of less than -128 were not correctly handled
* Problems with the solver under occasionally binding constraints
(occbin_solver command):
when solving the baseline regime, it would not properly handle
errors like
Blanchard-Kahn violationsthe piecewise linear Kalman filter (PKF)
would crash if the model solution
could not be computed for a parameter drawthe
oo_.FilteredVariablesKStepAhead and oo_.UpdatedVariables
MATLAB/Octave variables would contain the steady state twicethe
inversion filter would crash if the filter_step_ahead or
state_uncertainty options were requestedthe PKF would crash if
filter_step_ahead=1 was specifiedthe PKF would crash if the
state_uncertainty option was specified
together with the smoother_redux optionthe last regime before the
system is back to normal times in the
two-constraints case could be wrongly set, possibly leading to wrong
simulations, lack of convergence or crashes
* Problems with identification (identification command):
with prior_mc>1 specified, it would incorrectly display the share of
rank
deficient Jacobiansit would crash during plotting or displaying
identification strength when
the necessary identification criteria based on moments could not be
computed
* The plot_shock_decomposition command would crash if invalid field
names
were encountered
* The shock_decomposition command would not pass specified initial
dates to
generated plots
* Various pathological cases encountered in steady state finding could
lead to
a crash
* The solve_algo=0 option of the steady command would not honor tolx
* In the dynare_sensitivity command, stability mapping would not
correctly
honor the prior bounds
As a reminder, the list of new features introduced in versions 5.x can
be found in the release notes for 5.0.
On behalf of the Dynare Team,
--
Sébastien Villemot
Economist at CEPREMAP
Dynare developer
https://sebastien.villemot.name
Dear Dynare friends,
We are pleased to announce the release of Dynare 5.2. This maintenance
release fixes various bugs.
The Windows, macOS and source packages are already available for
download atthe Dynare website. All users are strongly encouraged to
upgrade.
This release is compatible with MATLAB versions ranging from 8.3
(R2014a) to 9.12 (R2022a), and with GNU Octave version 6.4.0 (under
Windows).
Note for macOS users with an Apple Silicon processor, and who are also
MATLAB users: the official MATLAB version for use on those processors
is still the Intel version (running through Rosetta 2), so the official
Dynare package for download on our website is built for Intel only.
However, since Mathworks has released a beta version of MATLAB for
Apple Silicon, we created a beta package of Dynare that you can try
with it. See this forum thread for more
details:https://forum.dynare.org/t/testers-wanted-release-of-dynare-5-x-bet…
Here is a list of the problems identified in version 5.1 and that have
been fixed in version 5.2:
* Problems with the steady_state operator:
if a steady_state operator contained an algebraic expression
appearing
multiple times in the model and sufficiently complex to trigger the
creation of a temporary term, then the result of the operator would
be
wrong (the operator was essentially ignored)if a steady_state
operator contained a call to an external function, then
the result of the operator would be wrong (the operator was
essentially
ignored). A proper fix to this problem would require substantial
architectural changes, so for now it is forbidden to use an external
function inside a steady_state operator
* Pruning in particle filtering at order 2 was not using the exact
same formula
as the original Kim et al. (2008) paper. A second-order term entered
the
cross-product between states and shocks, where it should have been a
first-order term. This however would not lead to explosive
trajectories in
practice
* The simul_replic option of the stoch_simul command would not store
the
binary file in the Output folder
* Problems with Ramsey policy (ramsey_model/ramsey_policy commands):
steady state files would not work when auxiliary variables included
Lagrange multipliersfor linear competitive equilibrium laws of
motion, welfare evaluated at
higher order was erroneously equated to steady state welfare
* The discretionary_policy command would not always correctly infer
the
number of instruments and equations, leading to spurious error
messages
* Perfect foresight simulations of purely forward or backward models
would
crash if complex numbers were encountered
* When using both block and bytecode options of the model block, if
the
model was such that a sufficiently complex algebraic expression
appeared both
in the residuals and in the derivatives, leading to the creation of
a
temporary term, then the results could be incorrect under some
circumstances
* When using the bytecode option of the model block, leads of more
than
+127 or lags of less than -128 were not correctly handled
* Problems with the solver under occasionally binding constraints
(occbin_solver command):
when solving the baseline regime, it would not properly handle
errors like
Blanchard-Kahn violationsthe piecewise linear Kalman filter (PKF)
would crash if the model solution
could not be computed for a parameter drawthe
oo_.FilteredVariablesKStepAhead and oo_.UpdatedVariables
MATLAB/Octave variables would contain the steady state twicethe
inversion filter would crash if the filter_step_ahead or
state_uncertainty options were requestedthe PKF would crash if
filter_step_ahead=1 was specifiedthe PKF would crash if the
state_uncertainty option was specified
together with the smoother_redux optionthe last regime before the
system is back to normal times in the
two-constraints case could be wrongly set, possibly leading to wrong
simulations, lack of convergence or crashes
* Problems with identification (identification command):
with prior_mc>1 specified, it would incorrectly display the share of
rank
deficient Jacobiansit would crash during plotting or displaying
identification strength when
the necessary identification criteria based on moments could not be
computed
* The plot_shock_decomposition command would crash if invalid field
names
were encountered
* The shock_decomposition command would not pass specified initial
dates to
generated plots
* Various pathological cases encountered in steady state finding could
lead to
a crash
* The solve_algo=0 option of the steady command would not honor tolx
* In the dynare_sensitivity command, stability mapping would not
correctly
honor the prior bounds
As a reminder, the list of new features introduced in versions 5.x can
be found in the release notes for 5.0.
On behalf of the Dynare Team,
--
Sébastien Villemot
Economist at CEPREMAP
Dynare developer
https://sebastien.villemot.name
Dear Dynare friends,
We are pleased to announce the release of Dynare 5.1.
This maintenance release fixes various bugs.
The Windows, macOS and source packages are already available for
download atthe Dynare website.
All users are strongly encouraged to upgrade.
This release is compatible with MATLAB versions ranging from 8.3
(R2014a) to 9.12 (R2022a), and with GNU Octave version 6.4.0 (under
Windows).
Here is a list of the problems identified in version 5.0 and that have
been fixed in version 5.1:
* Various problems with perfect foresight simulations in combination
with
block and/or bytecode options of the model block:
Simulation with bytecode and stack_solve_algo=4 could give incorrect
results if the model has a linear block of type “Solve two
boundaries
simple/complete”Simulation with bytecode and stack_solve_algo=1
could fail to convergeSimulation with block (but without bytecode)
and stack_solve_algo=1
gave wrong results in the last simulation period if the model has a
block
of type “Solve two boundaries simple/complete”Simulation with
bytecode and block would give incorrect results if the
model has a linear block of type “Solve forward
simple/complete”Simulation with block (but without bytecode) would
crash or give
incorrect results if the model has a block of type “Solve
forward/backward
simple/complete”Simulation with bytecode, block and
stack_solve_algo={0,1,4} would
crash or give incorrect results if the model has a block of type
“Solve
forward/backward complete”Simulation with block (but without
bytecode) gave incorrect results if
the model has a block of type “Solve backward
simple/complete”Simulation with block (with or without bytecode)
could give incorrect
results if the model has a nonlinear block of type “Solve
forward/backward
simple/complete”Simulation with bytecode, block and
stack_solve_algo=4 could give
incorrect results if the model has a block of type “Solve
backward/forward
simple/complete” that follows a block of type “Solve two boundaries”
(in
the sense of the dependency graph)The convergence criterion in
simulations with block (but without
bytecode) was incorrect: the value of the tolf option from the
steady
command was used instead of the value of tolf option from the
perfect_foresight_solver command
* Various problems with steady state computation in combination with
block
and/or bytecode options of the model block:
Steady state computation with bytecode and block could fail if some
equations are marked [static]Steady state computation with bytecode,
block and solve_algo ⩽ 4 or ⩾
9 could failSteady state computation with bytecode, block and
solve_algo=6 would
crash or give incorrect results if the model has a block of type
“Solve
forward/backward complete”
* The check command would crash or give incorrect results when using
the
block option of the model block and if the model has a block of type
“Solve backward complete”
* The static and incidence options of the model_info command did not
work
as documented in the reference manual
* Various problems with the method_of_moments command:
It would crash if no matched_moments block is presentIt would always
load the full range of the first Excel sheet instead of the
xls_range of the specified xls_sheetSMM would crash if a parameter
draw triggers an error during
additional_optimizer_steps = 13The debug option could not be passed
to the command
* In the estimation command, the scale_file field of the
posterior_sampler_options option did not correctly load the scale
* The moments_varendo option of the estimation command could crash for
large models
* The resid command would not show name tags when used in conjunction
with
the ramsey_model command
* Simulations with the occbin_solver command would not work if there
is only
a surprise shock in the first period
* The Liu & West auxiliary particle filter could enter infinite loops
As a reminder, the list of new features introduced in versions 5.x can
be found in the release notes for 5.0.
On behalf of the Dynare Team,
--
Sébastien Villemot
Economist at CEPREMAP
Dynare developer
https://sebastien.villemot.name
Dear Friends,
This year the Dynare Summer School will be hosted, from May 30 to June
3, 2022, by École Normale Supérieure (Campus Jourdan). The courses will
focus on simulation and estimation of DSGE models with Dynare.
More information is available on our website:
https://www.dynare.org/events/dynare-summer-school-2022-applications-open
Applications should be sent to school(a)dynare.org (with a CV and a recent
research paper attached as PDFs) no later than March 27, 2022.
Best regards,
Stéphane.
Dear Dynare friends,
We are pleased to announce the release of Dynare 5.0. This major
release adds new features and fixes various bugs.
The Windows, macOS and source packages are already available for
download at the Dynare website. All users are strongly encouraged to
upgrade.
This release is compatible with MATLAB versions ranging from 8.3
(R2014a) to 9.11 (R2021b), and with GNU Octave version 6.4.0 (under
Windows).
The new tools for semi-structural models and the improvements on the
nonlinear solvers were funded by the ECB. Special thanks to Nikola
Bokan (ECB) for his contributions and numerous bug reports and fixes.
Major user-visible changes * New routines for simulating semi-structural (backward) models where
some equations incorporate expectations based on future values of a
VAR or trend component model. See the var_model,
trend_component_model and var_expectation_model commands, and the
var_expectation operator.
* New routines for simulating semi-structural models where some
equations are specified using the polynomial adjustment costs (PAC)
approach, as in the FRB/US model (see Brayton et al., 2014 and
Brayton et al., 2000) and the ECB-BASE model (see Angelini et al.,
2019). The forward-looking terms of the PAC equations can be
computed either using a satellite VAR model, or using full model-
consistent expectations. See the pac_model command and
the pac_expectation operator.
* New Method of Moments toolbox that provides functionality to
estimate parameters by (i) Generalized Method of Moments (GMM) up to
3rd-order pruned perturbation approximation or (ii) Simulated Method
of Moments (SMM) up to any perturbation approximation order. The
toolbox is inspired by replication codes accompanying Andreasen et
al. (2018), Born and Pfeifer (2014), and Mutschler (2018). It is
accessible via the new method_of_moments command and the new
matched_moments block. Moreover, by default, a new non-linear least
squares optimizer based on lsqnonlin is used for minimizing the
method of moments objective function (available under
mode_compute=13). GMM can further benefit from using gradient-based
optimizers (usinganalytic_standard_errors option and/or passing
'Jacobian','on' to the optimization options) as the Jacobian of the
moment conditions can be computed analytically.
* Implementation of the Occbin algorithm by Guerrieri and Iacoviello
(2015), together with the inversion filter of Cuba-Borda, Guerrieri,
Iacoviello, and Zhong (2019) and the piecewise Kalman filter of
Giovannini, Pfeiffer, and Ratto (2021). It is available via the new
block occbin_constraints and the new commands occbin_setup,
occbin_solver, occbin_graph, and occbin_write_regimes.
* Stochastic simulations stoch_simul now supports theoretical moments
at order=3 with pruning.stoch_simul now reports second moments based
on the pruned state space if the pruning option is set (in previous
Dynare releases it would report a second-order accurate result based
on the linear solution).
* Estimation Performance optimization to pruned state space systems
and Lyapunov solvers.New option mh_posterior_mode_estimation to
estimation to perform mode-finding by running the MCMC.New
heteroskedastic filter and smoother, where shock standard errors
may unexpectedly change in every period. Triggered by
the heteroskedastic_filter option of the estimation command, and
configured via the heteroskedastic_shocks block.New option
mh_tune_guess for setting the initial value for mh_tune_jscale.New
option smoother_redux to estimation and calib_smoother to trigger
computing the Kalman smoother on a restricted state space instead of
the full one.New block filter_initial_state for setting the initial
condition of the Kalman filter/smoother.New option
mh_initialize_from_previous_mcmc to the estimation command that
allows to pick initial values for a new MCMC from a previous one.The
xls_sheet option of the estimation command now takes a quoted string
as value. The former unquoted syntax is still accepted, but no
longer recommended.New option particle_filter_options to set various
particle filter options.
* Perfect foresight and extended path New specialized algorithm in
perfect_foresight_solver to deal with purely static problems.The
debug option of perfect_foresight_solver provides debugging
information if the Jacobian is singular.In deterministic models
(perfect foresight or extended path), exogenous variables with
lead/lags are now replaced by auxiliary variables. This brings those
models in line with the transformation done on stochastic models.
However, note that the transformation is still not exactly the same
between the two classes of models, because there is no need to take
into account the Jensen inequality for the latter. In deterministic
models, there is a one-to-one mapping between exogenous with
lead/lags and auxiliaries, while in stochastic models, an auxiliary
endogenous may correspond to a more complex nonlinear expression.
* Optimal policy Several improvements to evaluate_planner_objective:
it now applies a consistent approximation order when doing the
computation;in addition to the conditional welfare, it now also
provides the unconditional welfare;in a stochastic context, it now
works with higher order approximation (only the conditional welfare
is available for order ⩾ 3);it now also works in a perfect foresight
context.discretionary_policy is now able to solve nonlinear models
(it will then use their first-order approximation, and the
analytical steady state must be provided).
* Identification New option schur_vec_tol to the identification
command, for setting the tolerance level used to find nonstationary
variables in the Schur decomposition of the transition matrix.The
identification command now supports optimal policy.
* Shock decomposition The fast_realtime option of the
realtime_shock_decomposition command now accepts a vector of
integers, which runs the smoother for all the specified data
vintages.
* Macro processor Macroprocessor variables can be defined without a
value (they are assigned integer 1).
* LaTeX and JSON outputs New nocommutativity option to the dynare
command. This option tells the preprocessor not to use the
commutativity of addition and multiplication when looking for common
subexpressions. As a consequence, when using this option, equations
in various outputs (LaTeX, JSON…) will appear as the user entered
them (without terms or factors swapped). Note that using this option
may have a performance impact on the preprocessing stage, though it
is likely to be small.Model-local variables are now substituted out
as part of the various model transformations. This means that they
will no longer appear in LaTeX or in JSON files (for the latter,
they are still visible with json=parse or json=check).
* Compilation of the model (use_dll option) Block decomposition
(option block of model) can now be used in conjunction with the
use_dll option.The use_dll option can now directly be given to the
dynare command.
* dseries classes Routines for converting between time series
frequencies (e.g. daily to monthly) have been added.dseries now
supports bi-annual and daily frequency data.dseries can now import
data from DBnomics, via the mdbnomics plugin. Note that this does
not yet work under Octave. For the time being, the DBnomics plugin
must be installed separately.
* Misc improvements The histval_file and initval_file commands have
been made more flexible and now have functionalities similar to the
datafile option of the estimation command.When using the loglinear
option, the output from Dynare now clearly shows that the results
reported concern the log of the original variable.Options block and
bytecode of model can now be used in conjunction with model-local
variables (variables declared with a pound-sign #).The model_info
command now prints the typology of endogenous variables for non-
block decomposed models.The total computing time of a run (in
seconds) is now saved to oo_.time.New notime option to the dynare
command, to disable the printing and the saving of the total
computing time.New parallel_use_psexec command-line Windows-specific
option for parallel local clusters: when true (the default), use
psexec to spawn processes; when false, use start.When compiling from
source, it is no longer necessary to pass the MATLAB_VERSION version
to the configure script; the version is now automatically detected.
Incompatible changes * Dynare will now generally save its output in the MODFILENAME/Output
folder (or the DIRNAME/Output folder if the dirname option was
specified) instead of the main directory. Most importantly, this
concerns the_results.mat and the _mode.mat files.
* The structure of the oo_.planner_objective field has been changed,
in relation to the improvements to evaluate_planner_objective.
* The preprocessor binary has been renamed to dynare-preprocessor, and
is now located in a dedicated preprocessor subdirectory.
* The dynare command no longer accepts output=dynamic and
output=first(these options actually had no effect).
* The minimal required MATLAB version is now R2014a (8.3).
* The 32-bit support has been dropped for Windows.
Bugs that were present in 4.6.4 and that have been fixed in 5.0 * Equations marked with static-tags were not detrended when a deflator
was specified
* Parallel execution of dsge_var estimation was broken
* The preprocessor would incorrectly simplify forward-looking constant
equations of the form x(+1)=0 to imply x=0
* Under some circumstances, the use of the model_local_variable
statement would lead to a crash of the preprocessor
* When using the block-option without bytecode the residuals of the
static model were incorrectly displayed
* When using k_order_solver, the simult_ function ignored requested
approximation orders that differed from the one used to compute the
decision rules
* Stochastic simulations of the k_order_solver without pruning
iterated on the policy function with a zero shock vector for the
first (non-endogenous) period
* estimation would ignore the mean of non-zero observables if the mean
was 0 for the initial parameter vector
* mode_check would crash if a parameter was estimated to be exactly 0
* load_mh_file would not be able to load the proposal density if the
previous run was done in parallel
* load_mh_file would not work with MCMC runs from Dynare versions
before 4.6.2
* ramsey_model would not correctly work with lmmcp
* ramsey_model would crash if a non-scalar error code was encountered
during steady state finding.
* Using undefined objects in the planner_objective function would
yield an erroneous error message about the objective containing
exogenous variables
* model_diagnostics did not correctly handle a previous loglinear
option
* solve_algo=3 (csolve) would ignore user-set maxit and tolf options
* The planner_objective values were not based on the correct
initialization of auxiliary variables (if any were present)
* The nostrict command line option was not ignoring unused endogenous
variables in initval, endval, and histval
* prior_posterior_statistics_core could crash for models with
eigenvalues very close to 1
* The display of the equation numbers in debug mode related to issues
in the Jacobian would not correctly take auxiliary equations into
account
* The resid command was not correctly taking auxiliary and missing
equations related to optimal policy (ramsey_model,
discretionary_policy) into account
* bytecode would lock the dynamic.bin file upon encountering an
exception, requiring a restart of MATLAB to be able to rerun the
file
* Estimation with the block model option would crash when calling the
block Kalman filter
* The block model option would crash if no initval statement was
present
* Having a variable with the same name as the mod-file present in the
base workspace would result in a crash
* oo_.FilteredVariablesKStepAheadVariances was wrongly computed in the
Kalman smoother based on the previous period forecast error variance
* Forecasts after estimation would not work if there were lagged
exogenous variables present
* Forecasts after estimation with MC would crash if measurement errors
were present
* Smoother results would be infinity for auxiliary variables
associated with lagged exogenous variables
* In rare cases, the posterior Kalman smoother could crash due to
previously accepted draws violating the Blanchard-Kahn conditions
when using an unrestricted state space
* perfect_foresight_solver would crash for purely static problems
* Monte Carlo sampling in identification would crash if the minimal
state space for the Komunjer and Ng test could not be computed
* Monte Carlo sampling in identification would skip the computation of
identification statistics for all subsequent parameter draws if an
error was triggered by one draw
* The --steps-option of Dynare++ was broken
* smoother2histval would crash if variable names were too similar
* smoother2histval was not keeping track of whether previously stored
results were generated with loglinear
* The initval_file option was not supporting Dynare’s translation of a
model into a one lead/lag-model via auxiliary variables
References * Andreasen et al. (2018): “The pruned state-space system for non-
linear DSGE models: Theory and empirical applications,” Review of
Economic Studies, 85(1), 1–49
* Angelini, Bokan, Christoffel, Ciccarelli and Zimic (2019):
“Introducing ECB-BASE: The blueprint the new ECB semi-structural
model for the euro area,” ECB Working Paper no. 2315
* Born and Pfeifer (2014): “Policy risk and the business cycle,”
Journal of Monetary Economics, 68, 68–85
* Brayton, Davis and Tulip (2000): “Polynomial adjustment costs in
FRB/US,” Unpublished manuscript
* Brayton, Laubach, and Reifschneider (2014): “The FRB/US Model: A
tool for macroeconomic policy analysis,” FEDS Notes. Washington:
Board of Governors of the Federal Reserve System,
https://doi.org/10.17016/2380-7172.0012
* Cuba-Borda, Guerrieri, Iacoviello, and Zhong (2019): “Likelihood
evaluation of models with occasionally binding constraints,” Journal
of Applied Econometrics, 34(7), 1073–1085
* Giovannini, Pfeiffer, and Ratto (2021): “Efficient and robust
inference of models with occasionally binding constraints,” Working
Paper 2021-03, Joint Research Centre, European Commission
* Guerrieri and Iacoviello (2015): “OccBin: A toolkit for solving
dynamic models with occasionally binding constraints easily,”
Journal of Monetary Economics, 70, 22–38
* Mutschler (2018): “Higher-order statistics for DSGE models,”
Econometrics and Statistics, 6(C), 44–56
On behalf of the Dynare Team,
--
Sébastien Villemot
Economist at CEPREMAP
Dynare developer
https://sebastien.villemot.name
Dear Friends,
This year the Dynare Summer School will take place from June 21-25,
2021.
Due to the coronavirus outbreak the school will take place online. The
courses will focus on the simulation and estimation of DSGE models with
Dynare.
More information is available on our website:
https://www.dynare.org/events/dynare-summer-school-2021-applications-open
If interested, you can apply by sending a CV and a recent research paper
(PDFs only) to school(a)dynare.org no later than May 15, 2021.
Best,
Stéphane
On behalf of the Dynare Team
Dear Dynare friends,
We are pleased to announce the release of Dynare 4.6.4.
This maintenance release fixes various bugs.
The Windows, macOS and source packages are already available for
download at the Dynare website.
All users are strongly encouraged to upgrade.
This release is compatible with MATLAB versions ranging from 7.9
(R2009b) to 9.10 (R2021a), and with GNU Octave version 6.2.0 (under
Windows).
Here is a list of the problems identified in version 4.6.3 and that
have been fixed in version 4.6.4:
Passing multiple shock values through a MATLAB/Octave vector in a
mshocks block would not work
The mode_compute=12 option was broken
The use_mh_covariance_matrix option was ignored
The load_mh_file option together with mh_replic=0 would not allow
computing moments_varendo for a different list of variables
The forecast option was ignored when using mode_compute=0 without
a mode-file to execute the smoother
The discretionary_policy command would crash in the presence of news
shocks
The ramsey_constraints block would crash if the constraints
contained defined parameters
Identification would display a wrong error message if a unit root was
present and diffuse_filter had been set
Particle filter estimation would crash if the initial state covariance
became singular for a draw
Particle filter estimation would crash if k_order_solver option
was specified with options_.particle.pruning
The initial state covariance in particle filter estimation could be
NaN when using nonlinear_filter_initialization=2
despite options_.particles.pruning=1
Output of smoother results when using particle filters would be based
on order=1
Output of moments_varendo results when using particle filters would
be based on order=1
When decreasing the order in .mod files, oo_.dr could contain
stale results from higher orders
Estimation results using the particle filter at order=3 would be
incorrect if the restricted state space differed from the unrestricted
one
The mode_compute=102 option (SOLVEOPT) could return with Inf instead
of the last feasible value
Using analytic_derivation for Bayesian estimation would result in
wrong results when the multivariate Kalman filter entered the steady
state stage
Using analytic_derivation for maximum likelihood estimation would
result in a crash
When using the Bayesian smoother with filtered_vars, the field
for Filtered_Variables_X_step_ahead used the length of vector instead
of the actual steps in filter_step_ahead
mode_compute=1,3 crashed when analytic_derivation was specified
mode_compute=1,3,102 did only allow for post-MATLAB 2016a option names
The cova_compute=0 option was not working with user-
defined MCMC_jumping_covariance
The mode_compute=1 option was not working with analytic_derivation
Not all commands were honouring the M_.dname folder when saving
LaTeX output of the simulated variance decomposition for observables
with measurement error could have a wrong variable label
As a reminder, the list of new features introduced in versions 4.6.x
can be found in the release notes for 4.6.0.
On behalf of the Dynare Team,
--
Sébastien Villemot
Economist at CEPREMAP
Dynare developer
https://sebastien.villemot.name
Dear Dynare users,
We are pleased to announce the release of Dynare 4.6.3.
This maintenance release fixes various bugs.
The Windows, macOS and source packages are already available for download at
the Dynare website.
All users are strongly encouraged to upgrade.
This release is compatible with MATLAB versions ranging from 7.9 (R2009b) to
9.9 (R2020b), and with GNU Octave versions 5.2.0 (under Windows) and 4.4.1
(under macOS).
Here is a list of the problems identified in version 4.6.2 and that have been
fixed in version 4.6.3:
* Using an unknown symbol in irf_shocks option of stoch_simul would lead to
a crash of the preprocessor
* discretionary_policy would crash in the presence of news shocks
* identification would crash for purely forward-looking models
* The endogenous_prior option did not properly handle missing observations
* The auxiliary particle filter with pruning and resampling would crash
* Initialization of the state variance for particle filters was buggy
* An @#else clause after an @#ifndef was not correctly interpreted
* An @#elseif clause after an @#ifdef or an @#ifndef was not correctly
interpreted
* Perfect foresight simulations of models with a single equation would crash
when using either the lmmcp option or the linear_approximation
* Inequality constraints on endogenous variables (when using the lmmcp
option) were not enforced on purely backward or purely forward models
* Perfect foresight simulations with bytecode and block options could crash
if there was a purely forward variable whose value in all periods could be
evaluated backward (typically a process of the form y=a*y(+1)+e)
* extended_path was broken with bytecode
* Under Windows, with Octave, the k-order perturbation and MS-SBVAR MEX files
could not be loaded
* On Fedora (and possibly other GNU/Linux distributions), compilation from
source would fail against Octave 5
As a reminder, the list of new features introduced in versions 4.6.x can be
found in the release notes for
4.6.0.
For the Dynare Team,
--
Sébastien Villemot
Economist at CEPREMAP
Dynare developer
https://sebastien.villemot.name