NB: All defaults set by initialize_ms_sbvar_options.m and passed to mex functions.
Miscellaneous Commands needed by Matlab to link to the Mex Functions (below)
Dynare Option and Argument |
Comments |
Default |
Matlab Variable Name |
Previous Dynare Option Name |
|||||
coefficients_prior_hyperparameters |
<<vec_value>> |
Sets the hyper parameters for the sbvar (mu in sz_prd.m), this should be a vector of doubles of length 6 |
[1.0 1.0 0.1 1.2 1.0 1.0] |
|
|
||||
freq |
<<integer>> |
quarters or months |
4 |
q_m |
freq |
||||
initial_year |
<<integer>> |
beginning of the year |
|
yrBin |
initial_year |
||||
initial_subperiod |
<<integer>> |
beginning of the quarter or month |
1 |
qmBin |
initial_subperiod |
||||
final_year |
<<integer>> |
final year |
|
yrFin |
final_year |
||||
final_subperiod |
<<integer>> |
final month or quarter |
4 |
qmFin |
final_subperiod |
||||
datafile |
<<string>> |
Load data |
|
|
data_file |
||||
nlags |
<<integer>> |
Number of lags |
1 |
lags |
nlags |
||||
cross_restrictions |
<<integer>> |
1: cross-A0-and-A+ restrictions |
0 |
indxc0Pres |
cross_restrictions |
||||
contemp_reduced_form |
<<integer>> |
1: contemporaneous recursive reduced form |
0 |
Rform |
comtemp_reduced_form |
||||
bayesian_prior |
<<integer>> |
1: Bayesian prior |
1 |
indxPrior |
bayesian_prior |
||||
alpha |
<<double>> |
Alpha value for squared time-varying structural shock lambda. |
1.0 |
galp |
alpha |
||||
beta |
<<double>> |
Beta value for squared time-varying structural shock lambda. |
1.0 |
gbeta |
beta |
||||
gsig2_lmd |
<<integer>> |
Case 4 (no state change across variables (i) and across lags (l); only one scale factor for all lagged variables change from state to state). Normal prior variance for structural shock lambda, one scale factor for all lagged D+ in a given equation. |
502 |
gsig2_lmd |
gsig2_lmd |
||||
gsig2_lmdm |
<<integer>> |
Case 3 (no state change across lags (l) but allows all variables for a given lag to switch states). Normal prior variance for different variables in lagged D+. |
502 |
gsig2_lmdm |
gsig2_lmdm |
Mex Functions
General Dynare instructions for MS-SBVAR
I suggest to use several main Dynare keywords rather than options in a single Dynare command ms-sbvar
Top Level Options (from switch_dw/state_space/sbvar/dw_sbvar_command_line.c):
Option |
Value |
Default |
Dynare Option |
Preproccessor Option |
Comments |
estimate |
-- |
-- |
options_.ms.estimate_msmodel |
ms_estimatation |
produces estimation - posterior mode or maximum likelihood |
simulate |
-- |
-- |
options_.ms.print_draws |
ms_simulation |
produces simulation run |
mdd |
-- |
-- |
options_.ms.compute_mdd |
ms_compute_mdd |
produces marginal data density from posterior draws |
probabilities |
-- |
-- |
options_.ms.compute_probabilities |
ms_compute_probabilities |
produces smoothed or filtered probabilities |
forecast |
-- |
-- |
options_.ms.forecast |
ms_forecast |
produces mean forecast |
ir |
-- |
-- |
options_.ms.irf |
ms_irf |
produces impulse responses for a given set of parameters with or without error bands and taking into account or not of parameter posterior uncertainty |
ms_variance_decomposition |
-- |
-- |
options_.ms.variance_decomposition |
ms_variance_decomposition |
produce variance decomposition |
Options Accepted by All Mex Functions (from switch_dw/state_space/sbvar/dw_sbvar_command_line.c):
Option |
Value |
Default |
Dynare Option |
Preproccessor Option |
Comments |
ft |
<<string>> |
-- |
options_.ms.file_tag |
file_tag |
The input file tag. Is used to create input file names if the fs or fp options are not present |
di |
<<string>> |
./ |
not supported |
-- |
input files are in specified directory |
do |
<<string>> |
./ |
not supported |
-- |
output files are in specified directory |
fs |
<<string>> |
|
options_.ms.initialization_file_tag |
initialization_file_tag |
specification file name |
fp |
<<string>> |
|
options_.ms.estimation_file_tag |
estimation_file_tag |
parameter file name |
ph |
<<string>> |
|
not supported |
-- |
header for the parameters |
MLE |
<<string>> |
|
not supported |
-- |
Determines if estimation finds the posterior mode or maximum likelihood. If ph or is not present, determines whether the input parameter header is "MLE: " or "Posterior mode: ". |
fto |
<<string>> |
|
options_.ms.output_file_tag |
output_file_tag |
The output file tag. Used to create output filenames. |
fo |
<<string>> |
|
not supported |
-- |
output file name |
pho |
<<string>> |
|
not supported |
-- |
parameter header used for output |
seed |
<<integer>> |
0 (i.e. random) |
options_.DynareRandomStreams.seed |
set by set_dynare_seed() in .mod file |
seed value for random number generation |
Mex Functions and Their Options
Create Init File Syntax (from switch_dw/state_space/sbvar/create_init_file.c):
Mex Calling structure: create_init_file <matlab filename> <markov filename> <file tag>
Option |
Value |
Default |
Dynare Option |
Preproccessor Option |
Comments |
|
<matlab filename> |
-- |
None. Set automatically in sz_prd.m |
-- |
matlab file name |
|
<markov filename> |
-- |
options_.ms.markov_file |
markov_file |
Markov file name |
|
<file tag> |
-- |
options_.ms.output_file_tag |
output_file_tag |
output file tag |
Estimation (from switch_dw/state_space/sbvar/sbvar_estimate.c):
Mex Calling structure: ms_sbvar_command_line(<<string>>)
Option |
Value |
Default |
Dynare Option |
Preproccessor Option |
Comments |
cb |
<<float>> |
1e-3 |
options_.ms.convergence_starting_value |
convergence_starting_value |
convergence criterion starting value |
ce |
<<float>> |
1e-6 |
options_.ms.convergence_ending_value |
convergence_ending_value |
convergence criterion ending value |
ci |
<<float>> |
0.1 |
options_.ms.convergence_increment_value |
convergence_increment_value |
convergence criterion increment value |
ib |
<<integer>> |
50 |
options_.ms.max_iterations_starting_value |
max_iterations_starting_value |
maximum interations starting value |
ii |
<<float>> |
2.0 |
options_.ms.max_iterations_increment_value |
max_iterations_increment_value |
maximum interations increment value |
mb |
<<integer>> |
100 |
options_.ms.max_block_iterations |
max_block_iterations |
maximum block iterations |
repeat_max |
<<integer>> |
10 |
options_.ms.max_repeated_optimization_runs |
max_repeated_optimization_runs |
maximum number of times to repeat csminwel runs |
repeat_tol_obj |
<<float>> |
0.1 |
options_.ms.function_convergence_criterion |
function_convergence_criterion |
convergence criterion for objective function |
repeat_tol_parms |
<<float>> |
0.1 |
options_.ms.parameter_convergence_criterion |
parameter_convergence_criterion |
convergence criterion for parameter values |
random |
<<integer>> |
5 |
options_.ms.number_of_large_perturbations |
number_of_large_perturbations |
number of large perturbations to make at each stage of the estimation process. |
random_small |
<<integer>> |
5 |
options_.ms.number_of_small_perturbations |
number_of_small_perturbations |
number of small perturbations to make after the large perturbations have stopped improving. |
random_small_ndraws |
<<integer>> |
1 |
options_.ms.number_of_posterior_draws_after_perturbation |
number_of_posterior_draws_after_perturbation |
number of consecutive posterior draws to make when producing a small perturbation |
random_max |
<<integer>> |
20 |
options_.ms.max_number_of_stages |
max_number_of_stages |
maximum number of stages allowed |
random_tol_obj |
<<float>> |
0.1 |
options_.ms.random_function_convergence_criterion |
random_function_convergence_criterion |
convergence criterion for objective function |
random_tol_parms |
<<float>> |
0.1 |
options_.ms.random_parameter_convergence_criterion |
random_parameter_convergence_criterion |
convergence criterion for parameter values |
Simulation (from switch_dw/state_space/sbvar/sbvar_simulate.c):
Mex Calling structure: ms_sbvar_command_line(<<string>>)
Option |
Value |
Default |
Dynare Option |
Preproccessor Option |
Comments |
ndraws |
<<integer>> |
10,000 |
options_.ms.mh_replic |
mh_replic |
Number of draws to save |
burnin |
<<integer>> |
0.1*ndraws |
options_.ms.drop |
drop |
Number of burn-in draws |
thin |
<<integer>> |
1 |
options_.ms.thinning_factor |
thinning_factor |
Thinning factor. Total number of draws made is thin*ndraws + burnin |
append |
-- |
off |
not supported |
-- |
Append simulation and info to existing files if they exist and use last parameter draw as initial value. |
mh |
<<integer>> |
30,000 |
options_.ms.adaptive_mh_draws |
adaptive_mh_draws |
Tuning period for Metropolis-Hasting draws |
flat |
-- |
off |
not supported |
-- |
Produce flat output file and header file |
nofree |
-- |
off |
not supported |
-- |
Do not produce free parameters file |
nd1 |
-- |
off |
not supported |
-- |
Normalize diagonal of A0 to one (flat output only) |
Marginal Data Density (from switch_dw/state_space/sbvar/sbvar_mdd.c):
Mex Calling structure: ms_sbvar_command_line(<<string>>)
Option |
Value |
Default |
Dynare Option |
Preproccessor Option |
Comments |
t |
<<string>> |
|
not supported |
-- |
filemname tag |
outtag |
<<string>> |
|
not supported |
-- |
filemname tag -- supersedes t and ft |
pf |
<<string>> |
simulation_<tag>.out |
options_.ms.mh_file |
simulation_file_tag |
posterior draws fileanme |
pt |
<<integer>> |
3 |
options_.ms.mdd_proposal_type(1) |
proposal_type (3-dim vector, see below) |
proposal type (1=gaussian, 2=power, 3=truncated power, 4=step, 5=truncated gaussian) |
d |
<<integer>> |
100,000 |
options_.ms.mdd_proposal_draws |
mdd_proposal_draws |
number proposal draws |
use_mode |
-- |
on |
not supported |
-- |
use posterior mode as center |
use_mean |
-- |
off |
options_.ms.mdd_use_mean_center |
mdd_use_mean_center |
use posterior mean as center |
of |
<<string>> |
mdd_<tag>.out |
not supported |
-- |
output filename |
l |
<<float>> |
0.1 |
options_.ms.mdd_proposal_type(2) |
|
Lower cutoff for in terms of probability. Used by truncated power, truncated gaussian, step, and uniform. |
h |
<<float>> |
0.9 |
options_.ms.mdd_proposal_type(3) |
|
Upper cutoff for in terms of probability. Used by power, truncated power, truncated gaussian, step, and uniform. |
graph |
-- |
off |
not supported |
-- |
Produce output suitable for graphing the marginal cummulative densities of the radius of the posterior and elliptical distributions. The marginal data density is not computed. |
-- |
off |
not supported |
-- |
Prints the matrices posterior and proposal |
|
append |
-- |
off |
not supported |
-- |
Appends proposal draws to those previously made. If the posterior draws have changed in any way, then this option must not be used. The filename for the proposal draws is proposal_t<proposal type>_<tag>.out. |
Regime Probability (from switch_dw/state_space/sbvar/sbvar_probabilities.c):
Mex Calling structure: ms_sbvar_command_line(<<string>>)
Option |
Value |
Default |
Dynare Option |
Preproccessor Option |
Comments |
filtered |
-- |
off |
options_.ms.filtered_probabilities |
filtered_probabilities |
filtered probabilities are computed instead of smoothed |
real_time_smoothed |
-- |
off |
options_.ms.real_time_smoothed_probabilities |
real_time_smoothed_probabilities |
smoothed probabilities are computed based on time t information are computed for 0 <= t <= nobs |
Impulse Responses (from switch_dw/state_space/sbvar/sbvar_impulse_responses.c):
Mex Calling structure: ms_sbvar_command_line(<<string>>)
Option |
Value |
Default |
Dynare Option |
Preproccessor Option |
Comments |
horizon |
<<integer>> |
12 |
options_.ms.horizon |
horizon |
forecast horizon |
filtered |
-- |
off |
options_.ms.filtered_probabilities |
filtered_probabilities |
Uses filtered probabilities at the end of the sample as initial conditions for regime probabilities. This flag only applies if neither regimes nor regime is specified. |
percentiles |
<<vector>> |
if error_bands, [0.16 0.50 0.84] else [0.50] |
options_.ms.error_band_percentiles |
error_band_percentiles |
Percentiles to compute. The first parameter after percentiles must be the number of percentiles and the following values are the actual percentiles. |
parameter_uncertainty |
-- |
off |
options_.ms.parameter_uncertainty |
parameter_uncertainty |
Apply parameter uncertainty when computing percentiles. |
shocks_per_parameter |
<<integer>> |
10,000 or 10 |
options_.ms.shocks_per_parameter & options_.ms.shock_draws |
shock_draws |
Number of regime paths to draw for each parameter draw. The default value is 1 if parameter_uncertainty is set and 10,000 otherwise. |
thin |
<<integer>> |
1 |
options_.ms.thinning_factor |
thinning_factor |
Thinning factor. Only 1/thin of the draws in posterior draws file are used. |
regimes |
-- |
off |
options_.ms.regimes |
regimes |
Describes the evolution of regimes. |
regime |
<<integer>> |
off |
options_.ms.regime |
regime |
Given the data and model parameters, what is the ergodic probability of being in the specified regime. |
flat |
-- |
off |
not supported |
-- |
Assumes that posterior draws is a flat file format |
median |
-- |
off |
options_.ms.median |
median |
This is intended as a shortcut to setting percentiles as percentiles=[0.50] |
Forecast (from switch_dw/state_space/sbvar/sbvar_forecast.c):
Mex Calling structure: ms_sbvar_command_line(<<string>>)
Option |
Value |
Default |
Dynare Option |
Preproccessor Option |
Comments |
horizon |
<<integer>> |
12 |
options_.ms.horizon |
horizon |
See horizon in irf |
number_observations |
<<integer>> |
0 |
options_.ms.forecast_data_obs |
data_obs_nbr |
the number of data points included in the output |
percentiles |
<<vector>> |
if error_bands, [0.16 0.50 0.84] else [0.50] |
options_.ms.error_band_percentiles |
error_band_percentiles |
See percentiles in irf |
parameter_uncertainty |
-- |
off |
options_.ms.parameter_uncertainty |
parameter_uncertainty |
See parameter_uncertainty in irf |
shocks_per_parameter |
<<integer>> |
10,000 or 10 |
options_.ms.shocks_per_parameter & options_.ms.shock_draws |
shock_draws |
See shocks_per_parameter in irf |
thin |
<<integer>> |
1 |
options_.ms.thinning_factor |
thinning_factor |
See thin in irf |
regimes |
<<integer>> |
0 |
options_.ms.regimes |
regimes |
See regimes in irf |
regime |
<<integer>> |
0 |
options_.ms.regime |
regime |
See regime in irf |
mean |
-- |
off |
not supported in Dan's code |
-- |
produces mean forecast |
flat |
-- |
off |
not supported |
-- |
See flat in irf |
nodate |
-- |
off |
not supported |
-- |
If not defined, then the first column of the output is a date number. A date number of one is the first date in the data series, a date number of nobs is the last date in the data series, and date numbers larger than this correspond to forecasts. |
median |
-- |
off |
options_.ms.median |
median |
See analogous table under irf |
Variance Decomposition (from switch_dw/state_space/sbvar/sbvar_variance_decomposition.c):
Mex Calling structure: ms_sbvar_command_line(<<string>>)
Option |
Value |
Default |
Dynare Option |
Preproccessor Option |
Comments |
horizon |
<<integer>> |
12 |
options_.ms.horizon |
horizon |
forecast horizon |
filtered |
-- |
off |
options_.ms.filtered_probabilities |
filtered_probabilities |
See filtered in irf |
error_bands |
-- |
on |
options_.ms.error_bands |
no_error_bands |
Calculate the mean |
percentiles |
<<vector>> |
|
options_.ms.error_band_percentiles |
error_band_percentiles |
See percentiles in irf |
parameter_uncertainty |
-- |
off |
options_.ms.parameter_uncertainty |
parameter_uncertainty |
See parameter_uncertainty in irf |
shocks_per_parameter |
<<integer>> |
10,000 or 10 |
options_.ms.shocks_per_parameter & options_.ms.shock_draws |
shock_draws |
See shocks_per_parameter in irf |
thin |
<<integer>> |
1 |
options_.ms.thinning_factor |
thinning_factor |
See thin in irf |
regimes |
-- |
off |
options_.ms.regimes |
regimes |
See regimes in irf |
regime |
<<integer>> |
0 |
options_.ms.regime |
regime |
See regime in irf |
flat |
-- |
off |
not supported |
-- |
See flat in irf |