From d8424b162073d51a58030787ea5b74b76e544d6e Mon Sep 17 00:00:00 2001 From: Marco Ratto <marco.ratto@jrc.ec.europa.eu> Date: Mon, 20 Apr 2015 11:06:21 +0200 Subject: [PATCH] updated documentation for IRF/moment restrictions and refreshed sensitivity analyses and plots --- doc/dynare.texi | 794 +++++++++++++++++++++++++++++------------------- 1 file changed, 484 insertions(+), 310 deletions(-) diff --git a/doc/dynare.texi b/doc/dynare.texi index fcb976aecd..05c58ce42e 100644 --- a/doc/dynare.texi +++ b/doc/dynare.texi @@ -241,13 +241,19 @@ Stochastic solution and simulation Sensitivity and identification analysis +* Performing sensitivity analysis:: +* Performing identification analysis:: +* IRF/moment calibration:: +* Types of analysis and output files:: + +Types of analysis and output files * Sampling:: * Stability Mapping:: +* IRF/moment restrictions:: * Reduced Form Mapping:: * RMSE:: * Screening Analysis:: * Identification Analysis:: -* Performing Sensitivity and Identification Analysis:: Macro-processing language @@ -6869,311 +6875,8 @@ With respect to the previous version of the toolbox, in order to work properly, the GSA toolbox no longer requires that the Dynare estimation environment is set up. -Sensitivity analysis results are saved locally in @code{<mod_file>/GSA}, -where @code{<mod_file>.mod} is the name of the DYNARE model file. - -@menu -* Sampling:: -* Stability Mapping:: -* Reduced Form Mapping:: -* RMSE:: -* Screening Analysis:: -* Identification Analysis:: -* Performing Sensitivity and Identification Analysis:: -@end menu - -@node Sampling -@subsection Sampling - -The following binary files are produced: -@itemize -@item -@code{<mod_file>_prior.mat}: this file stores information about the analyses -performed sampling from the prior ranges, @i{i.e.} @code{pprior=1} and @code{ppost=0}; - -@item -@code{<mod_file>_mc.mat}: this file stores information about the analyses performed -sampling from multivariate normal, @i{i.e.} @code{pprior=0} and @code{ppost=0}; - -@item -@code{<mod_file>_post.mat}: this file stores information about analyses performed -using the Metropolis posterior sample, @i{i.e.} @code{ppost=1}. -@end itemize - -@node Stability Mapping -@subsection Stability Mapping - -Figure files produced are of the form @code{<mod_file>_prior_*.fig} and store results -for stability mapping from prior Monte-Carlo samples: -@itemize -@item -@code{<mod_file>_prior_stab_SA_*.fig}: plots of the Smirnov test analyses -confronting the cdf of the sample fulfilling Blanchard-Kahn conditions -with the cdf of the rest of the sample; - -@item -@code{<mod_file>_prior_stab_indet_SA_*.fig}: plots of the Smirnov test -analyses confronting the cdf of the sample producing indeterminacy -with the cdf of the original prior sample; - -@item -@code{<mod_file>_prior_stab_unst_SA_*.fig}: plots of the Smirnov test -analyses confronting the cdf of the sample producing unstable (explosive -roots) behavior with the cdf of the original prior sample; - -@item -@code{<mod_file>_prior_stable_corr_*.fig}: plots of bivariate projections -of the sample fulfilling Blanchard-Kahn conditions; - -@item -@code{<mod_file>_prior_indeterm_corr_*.fig}: plots of bivariate projections -of the sample producing indeterminacy; - -@item -@code{<mod_file>_prior_unstable_corr_*.fig}: plots of bivariate projections -of the sample producing instability; - -@item -@code{<mod_file>_prior_unacceptable_corr_*.fig}: plots of bivariate projections -of the sample producing unacceptable solutions, @i{i.e.} either -instability or indeterminacy or the solution could not be found (@i{e.g.} -the steady state solution could not be found by the solver). -@end itemize - -Similar conventions apply for @code{<mod_file>_mc_*.fig} files, obtained when -samples from multivariate normal are used. - -@node Reduced Form Mapping -@subsection Reduced Form Mapping - -The mapping of the reduced form solution forces the use of samples from -prior ranges or prior distributions, @i{i.e.}: @code{pprior=1} and @code{ppost=0}. It -uses 250 samples to optimize smoothing parameters and 1000 samples to compute the -fit. The rest of the sample is used for out-of-sample validation. One can also -load a previously estimated mapping with a new Monte-Carlo sample, to look at the -forecast for the new Monte-Carlo sample. - -The following synthetic figures are produced: -@itemize -@item -@code{<mod_file>_redform_<endo name>_vs_lags_*.fig}: shows bar charts -of the sensitivity indices for the ten most important parameters driving -the reduced form coefficients of the selected endogenous variables -(@code{namendo}) versus lagged endogenous variables (@code{namlagendo}); suffix -@code{log} indicates the results for log-transformed entries; - -@item -@code{<mod_file>_redform_<endo name>_vs_shocks_*.fig}: shows bar charts -of the sensitivity indices for the ten most important parameters driving -the reduced form coefficients of the selected endogenous variables -(@code{namendo}) versus exogenous variables (@code{namexo}); suffix @code{log} -indicates the results for log-transformed entries; - -@item -@code{<mod_file>_redform_GSA(_log).fig}: shows bar chart of all sensitivity -indices for each parameter: this allows one to notice parameters that -have a minor effect for any of the reduced form coefficients. -@end itemize - -Detailed results of the analyses are shown in the subfolder @code{<mod_file>/GSA/redform_stab}, -where the detailed results of the estimation of the single functional relationships -between parameters @math{\theta} and reduced form coefficient are stored in separate directories -named as: - -@itemize -@item -@code{<namendo>_vs_<namlagendo>}: for the entries of the transition matrix; - -@item -@code{<namendo>_vs_<namexo>}: for entries of the matrix of the shocks. -@end itemize -Moreover, analyses for log-transformed entries are denoted with the following -suffixes (@math{y} denotes the generic reduced form coefficient): -@itemize -@item -@code{log}: @math{y^* = \log(y)}; -@item -@code{minuslog}: @math{y^* = \log(-y)}; -@item -@code{logsquared}: @math{y^* = \log(y^2)} for symmetric fat tails; -@item -@code{logskew}: @math{y^* = \log(|y + \lambda|)} for asymmetric fat tails. -@end itemize -The optimal type of transformation is automatically selected without the -need of user intervention. - -@node RMSE -@subsection RMSE - -The RMSE analysis can be performed with different types of sampling options: -@enumerate -@item -When @code{pprior=1} and @code{ppost=0}, the toolbox analyzes the RMSEs for -the Monte-Carlo sample obtained by sampling parameters from their prior distributions -(or prior ranges): this analysis provides some hints about -what parameter drives the fit of which observed series, prior to the full -estimation; - -@item -When @code{pprior=0} and @code{ppost=0}, the toolbox analyzes the RMSEs for -a multivariate normal Monte-Carlo sample, with covariance matrix based on -the inverse Hessian at the optimum: this analysis is useful when maximum likelihood -estimation is done (@i{i.e.} no Bayesian estimation); - -@item -When @code{ppost=1} the toolbox analyzes the RMSEs for the posterior sample -obtained by Dynare's Metropolis procedure. -@end enumerate - -The use of cases 2 and 3 requires an estimation step beforehand. To -facilitate the sensitivity analysis after estimation, the @code{dynare_sensitivity} -command also allows you to indicate some options of the @code{estimation} -command. These are: -@itemize @bullet -@item @code{datafile} -@item @code{nobs} -@item @code{first_obs} -@item @code{prefilter} -@item @code{presample} -@item @code{nograph} -@item @code{nodisplay} -@item @code{graph_format} -@item @code{conf_sig} -@item @code{loglinear} -@item @code{mode_file} -@end itemize - -Binary files produced my RMSE analysis are: -@itemize -@item -@code{<mod_file>_prior_*.mat}: these files store the filtered and smoothed - variables for the prior Monte-Carlo sample, generated when doing RMSE analysis - (@code{pprior=1} and @code{ppost=0}); -@item -@code{<mode_file>_mc_*.mat}: these files store the filtered and smoothed variables - for the multivariate normal Monte-Carlo sample, generated when doing - RMSE analysis (@code{pprior=0} and @code{ppost=0}). -@end itemize - -Figure files @code{<mod_file>_rmse_*.fig} store results for the RMSE analysis. - -@itemize -@item -@code{<mod_file>_rmse_prior*.fig}: save results for the analysis using prior -Monte-Carlo samples; - -@item -@code{<mod_file>_rmse_mc*.fig}: save results for the analysis using multivariate -normal Monte-Carlo samples; - -@item -@code{<mod_file>_rmse_post*.fig}: save results for the analysis using Metropolis -posterior samples. -@end itemize - -The following types of figures are saved (we show prior sample to fix ideas, -but the same conventions are used for multivariate normal and posterior): - -@itemize -@item -@code{<mod_file>_rmse_prior_*.fig}: for each parameter, plots the cdfs -corresponding to the best 10% RMSEs of each observed series; - -@item -@code{<mod_file>_rmse_prior_dens_*.fig}: for each parameter, plots the -pdfs corresponding to the best 10% RMESs of each observed series; - -@item -@code{<mod_file>_rmse_prior_<name of observedseries>_corr_*.fig}: for -each observed series plots the bi-dimensional projections of samples -with the best 10% RMSEs, when the correlation is significant; - -@item -@code{<mod_file>_rmse_prior_lnlik*.fig}: for each observed series, plots -in red the cdf of the log-likelihood corresponding to the best 10% -RMSEs, in green the cdf of the rest of the sample and in blue the -cdf of the full sample; this allows one to see the presence of some -idiosyncratic behavior; - -@item -@code{<mod_file>_rmse_prior_lnpost*.fig}: for each observed series, plots -in red the cdf of the log-posterior corresponding to the best 10% RMSEs, -in green the cdf of the rest of the sample and in blue the cdf of the full -sample; this allows one to see idiosyncratic behavior; - -@item -@code{<mod_file>_rmse_prior_lnprior*.fig}: for each observed series, plots -in red the cdf of the log-prior corresponding to the best 10% RMSEs, -in green the cdf of the rest of the sample and in blue the cdf of the full -sample; this allows one to see idiosyncratic behavior; - -@item -@code{<mod_file>_rmse_prior_lik_SA_*.fig}: when @code{lik_only=1}, this shows -the Smirnov tests for the filtering of the best 10% log-likelihood values; - -@item -@code{<mod_file>_rmse_prior_post_SA_*.fig}: when @code{lik_only=1}, this shows -the Smirnov test for the filtering of the best 10% log-posterior values. -@end itemize - -@node Screening Analysis -@subsection Screening Analysis - -Screening analysis does not require any additional options with respect to -those listed in @ref{Sampling Options}. The toolbox performs all the -analyses required and displays results. - -The results of the screening analysis with Morris sampling design are stored -in the subfolder @code{<mod_file>/GSA/SCREEN}. The data file @code{<mod_file>_prior} stores -all the information of the analysis (Morris sample, reduced form coefficients, -etc.). - -Screening analysis merely concerns reduced form coefficients. Similar -synthetic bar charts as for the reduced form analysis with Monte-Carlo samples are -saved: -@itemize -@item -@code{<mod_file>_redform_<endo name>_vs_lags_*.fig}: shows bar charts -of the elementary effect tests for the ten most important parameters -driving the reduced form coefficients of the selected endogenous variables -(@code{namendo}) versus lagged endogenous variables (@code{namlagendo}); - -@item -@code{<mod_file>_redform_<endo name>_vs_shocks_*.fig}: shows bar charts -of the elementary effect tests for the ten most important parameters -driving the reduced form coefficients of the selected endogenous variables -(@code{namendo}) versus exogenous variables (@code{namexo}); - -@item -@code{<mod_file>_redform_screen.fig}: shows bar chart of all elementary -effect tests for each parameter: this allows one to identify parameters that -have a minor effect for any of the reduced form coefficients. -@end itemize - -@node Identification Analysis -@subsection Identification Analysis - -Setting the option @code{identification=1}, an identification analysis based on -theoretical moments is performed. Sensitivity plots are provided that allow -to infer which parameters are most likely to be less identifiable. - -Prerequisite for properly running all the identification routines, is the keyword -@code{identification}; in the Dynare model file. This keyword triggers -the computation of analytic derivatives of the model with respect to estimated -parameters and shocks. This is required for option @code{morris=2}, -which implements @cite{Iskrev (2010)} identification analysis. - -For example, the placing @code{identification; dynare_sensitivity(identification=1, morris=2);} -in the Dynare model file trigger identification analysis using analytic derivatives -@cite{Iskrev (2010)}, jointly with the mapping of the acceptable region. - -The identification analysis with derivatives can also be triggered by the -commands @code{identification;} This does not do the mapping of -acceptable regions for the model and uses the standard random sampler of Dynare. -It completely offsets any use of the sensitivity analysis toolbox. - -@node Performing Sensitivity and Identification Analysis -@subsection Performing Sensitivity and Identification Analysis +@node Performing sensitivity analysis +@subsection Performing sensitivity analysis @deffn Command dynare_sensitivity ; @deffnx Command dynare_sensitivity (@var{OPTIONS}@dots{}); @@ -7406,12 +7109,86 @@ Maximum number of lags for moments in identification analysis. Default: @code{1} @end deffn -@deffn Command identification ; -@deffnx Command identification (@var{OPTIONS}@dots{}); +@node IRF/Moment calibration +@subsection IRF/Moment calibration -@descriptionhead +IRF and moment calibration can be defined in @code{irf_calibration} and @code{moment_calibration} blocks: -This command triggers identification analysis. +@deffn Block irf_calibration; +@deffnx Block irf_calibration(@var{OPTIONS}@dots{}); + +@descriptionhead + +This block allows defining IRF calibration criteria and is terminated by @code{end;}. +To set IRF sign restrictions, the following syntax is used +@example +@var{VARIABLE_NAME}(@var{INTEGER}),@var{EXOGENOUS_NAME}, -; +@var{VARIABLE_NAME}(@var{INTEGER}:@var{INTEGER}),@var{EXOGENOUS_NAME}, +; +@end example +To set IRF restrictions with specific intervals, the following syntax is used +@example +@var{VARIABLE_NAME}(@var{INTEGER}),@var{EXOGENOUS_NAME}, [@var{DOUBLE} @var{DOUBLE}]; +@var{VARIABLE_NAME}(@var{INTEGER}:@var{INTEGER}),@var{EXOGENOUS_NAME}, [@var{DOUBLE} @var{DOUBLE}]; +@end example + +When @code{(@var{INTEGER}:@var{INTEGER})} is used, the restriction is considered to be fulfilled by a logical OR. +A list of restrictions must always be fulfilled with logical AND. + +@examplehead + +@example +irf_calibration; +y(1:4), e_ys, [ -50 50]; //[first year response with logical OR] +@@#for ilag in 21:40 +R_obs(@@@{ilag@}), e_ys, [0 6]; //[response from 5th to 10th years with logical AND] +@@#endfor +end; +@end example + +@end deffn + +@deffn Block moment_calibration; +@deffnx Block moment_calibration(@var{OPTIONS}@dots{}); + +@descriptionhead + +This block allows defining moment calibration criteria. This block is terminated by @code{end;}, and contains lines of the +form: +@example +@var{VARIABLE_NAME1},@var{VARIABLE_NAME2}(+/-@var{INTEGER}), [@var{DOUBLE} @var{DOUBLE}]; +@var{VARIABLE_NAME1},@var{VARIABLE_NAME2}(+/-@var{INTEGER}), +/-; +@var{VARIABLE_NAME1},@var{VARIABLE_NAME2}(+/-(@var{INTEGER}:@var{INTEGER})), [@var{DOUBLE} @var{DOUBLE}]; +@var{VARIABLE_NAME1},@var{VARIABLE_NAME2}((-@var{INTEGER}:+@var{INTEGER})), [@var{DOUBLE} @var{DOUBLE}]; +@end example + +When @code{(@var{INTEGER}:@var{INTEGER})} is used, the restriction is considered to be fulfilled by a logical OR. +A list of restrictions must always be fulfilled with logical AND. + +@examplehead + +@example +moment_calibration; +y_obs,y_obs, [0.5 1.5]; //[unconditional variance] +y_obs,y_obs(-(1:4)), +; //[sign restriction for first year acf with logical OR] +@@#for ilag in -2:2 +y_obs,R_obs(@@@{ilag@}), -; //[-2:2 ccf with logical AND] +@@#endfor +@@#for ilag in -4:4 +y_obs,pie_obs(@@@{ilag@}), -; //[-4_4 ccf with logical AND] +@@#endfor +end; +@end example + +@end deffn +@node Performing identification analysis +@subsection Performing identification analysis + +@deffn Command identification ; +@deffnx Command identification (@var{OPTIONS}@dots{}); + +@descriptionhead + +This command triggers identification analysis. @optionshead @@ -7486,6 +7263,403 @@ Specify the parameter set to use. Default: @code{prior_mean} @end deffn +@node Types of analysis and output files +@subsection Types of analysis and output files + +The sensitivity analysis toolbox includes several types of analyses. +Sensitivity analysis results are saved locally in @code{<mod_file>/GSA}, +where @code{<mod_file>.mod} is the name of the DYNARE model file. + +@menu +* Sampling:: +* Stability Mapping:: +* IRF/moment restrictions:: +* Reduced Form Mapping:: +* RMSE:: +* Screening Analysis:: +* Identification Analysis:: +@end menu + +@node Sampling +@subsubsection Sampling + +The following binary files are produced: +@itemize +@item +@code{<mod_file>_prior.mat}: this file stores information about the analyses +performed sampling from the prior, @i{i.e.} @code{pprior=1} and @code{ppost=0}; + +@item +@code{<mod_file>_mc.mat}: this file stores information about the analyses performed +sampling from multivariate normal, @i{i.e.} @code{pprior=0} and @code{ppost=0}; + +@item +@code{<mod_file>_post.mat}: this file stores information about analyses performed +using the Metropolis posterior sample, @i{i.e.} @code{ppost=1}. +@end itemize + +@node Stability Mapping +@subsubsection Stability Mapping + +Figure files produced are of the form @code{<mod_file>_prior_*.fig} and store results +for stability mapping from prior Monte-Carlo samples: +@itemize +@item +@code{<mod_file>_prior_stable.fig}: plots of the Smirnov test and the correlation analyses +confronting the cdf of the sample fulfilling Blanchard-Kahn conditions (blue color) +with the cdf of the rest of the sample (red color), @i{i.e.} either +instability or indeterminacy or the solution could not be found (@i{e.g.} +the steady state solution could not be found by the solver); + +@item +@code{<mod_file>_prior_indeterm.fig}: plots of the Smirnov test and the correlation +analyses confronting the cdf of the sample producing indeterminacy (red color) +with the cdf of the rest of the sample (blue color); + +@item +@code{<mod_file>_prior_unstable.fig}: plots of the Smirnov test and the correlation +analyses confronting the cdf of the sample producing explosive roots (red color) +with the cdf of the rest of the sample (blue color); + +@item +@code{<mod_file>_prior_wrong.fig}: plots of the Smirnov test and the correlation +analyses confronting the cdf of the sample where the solution could not be found (@i{e.g.} +the steady state solution could not be found by the solver - red color) +with the cdf of the rest of the sample (blue color); + +@item +@code{<mod_file>_prior_calib.fig}: plots of the Smirnov test and the correlation +analyses splitting the sample fulfilling Blanchard-Kahn conditions, +by confronting the cdf of the sample where IRF/moment restrictions are matched (blue color) +with the cdf where IRF/moment restrictions are NOT matched (red color); + +@end itemize + +Similar conventions apply for @code{<mod_file>_mc_*.fig} files, obtained when +samples from multivariate normal are used. + +@node IRF/Moment restrictions +@subsubsection IRF/Moment restrictions + +The following binary files are produced: +@itemize +@item +@code{<mod_file>_prior_restrictions.mat}: this file stores information about the IRF/moment restriction analysis +performed sampling from the prior ranges, @i{i.e.} @code{pprior=1} and @code{ppost=0}; + +@item +@code{<mod_file>_mc_restrictions.mat}: this file stores information about the IRF/moment restriction analysis performed +sampling from multivariate normal, @i{i.e.} @code{pprior=0} and @code{ppost=0}; + +@item +@code{<mod_file>_post_restrictions.mat}: this file stores information about IRF/moment restriction analysis performed +using the Metropolis posterior sample, @i{i.e.} @code{ppost=1}. +@end itemize + +Figure files produced are of the form @code{<mod_file>_prior_irf_calib_*.fig} and @code{<mod_file>_prior_moment_calib_*.fig} and store results +for mapping restrictions from prior Monte-Carlo samples: +@itemize +@item +@code{<mod_file>_prior_irf_calib_<ENDO_NAME>_vs_<EXO_NAME>_<PERIOD>.fig}: plots of the Smirnov test and the correlation +analyses splitting the sample fulfilling Blanchard-Kahn conditions, +by confronting the cdf of the sample where the individual IRF restriction +@code{<ENDO_NAME>} vs. @code{<EXO_NAME>} at period(s) @code{<PERIOD>} is matched (blue color) +with the cdf where the IRF restriction is NOT matched (red color) + +@item +@code{<mod_file>_prior_irf_calib_<ENDO_NAME>_vs_<EXO_NAME>_ALL.fig}: plots of the Smirnov test and the correlation +analyses splitting the sample fulfilling Blanchard-Kahn conditions, +by confronting the cdf of the sample where ALL the individual IRF restrictions for the same couple +@code{<ENDO_NAME>} vs. @code{<EXO_NAME>} are matched (blue color) +with the cdf where the IRF restriction is NOT matched (red color) + +@item +@code{<mod_file>_prior_irf_restrictions.fig}: plots visual information on the IRF restrictions +compared to the actual Monte Carlo realization from prior sample. + +@item +@code{<mod_file>_prior_moment_calib_<ENDO_NAME1>_vs_<ENDO_NAME2>_<LAG>.fig}: plots of the Smirnov test and the correlation +analyses splitting the sample fulfilling Blanchard-Kahn conditions, +by confronting the cdf of the sample where the individual acf/ccf moment restriction +@code{<ENDO_NAME1>} vs. @code{<ENDO_NAME2>} at lag(s) @code{<LAG>} is matched (blue color) +with the cdf where the IRF restriction is NOT matched (red color) + +@item +@code{<mod_file>_prior_moment_calib_<ENDO_NAME>_vs_<EXO_NAME>_ALL.fig}: plots of the Smirnov test and the correlation +analyses splitting the sample fulfilling Blanchard-Kahn conditions, +by confronting the cdf of the sample where ALL the individual acf/ccf moment restrictions for the same couple +@code{<ENDO_NAME1>} vs. @code{<ENDO_NAME2>} are matched (blue color) +with the cdf where the IRF restriction is NOT matched (red color) + +@item +@code{<mod_file>_prior_moment_restrictions.fig}: plots visual information on the moment restrictions +compared to the actual Monte Carlo realization from prior sample. + +@end itemize + +Similar conventions apply for @code{<mod_file>_mc_*.fig} and @code{<mod_file>_post_*.fig} files, obtained when +samples from multivariate normal or from posterior are used. + +@node Reduced Form Mapping +@subsubsection Reduced Form Mapping + +When the option @code{threshold_redform} is not set, or it is empty (the default), this analysis estimates a multivariate +smoothing spline ANOVA model (the 'mapping') for the selected entries in the transition matrix of the shock matrix of the reduce form first order solution of the model. +The mapping of the reduced form solution forces the use of samples from +prior ranges or prior distributions, @i{i.e.}: @code{pprior=1} and @code{ppost=0}. It +uses 250 samples to optimize smoothing parameters and 1000 samples to compute the +fit. The rest of the sample is used for out-of-sample validation. One can also +load a previously estimated mapping with a new Monte-Carlo sample, to look at the +forecast for the new Monte-Carlo sample. + +The following synthetic figures are produced: +@itemize +@item +@code{<mod_file>_redform_<endo name>_vs_lags_*.fig}: shows bar charts +of the sensitivity indices for the ten most important parameters driving +the reduced form coefficients of the selected endogenous variables +(@code{namendo}) versus lagged endogenous variables (@code{namlagendo}); suffix +@code{log} indicates the results for log-transformed entries; + +@item +@code{<mod_file>_redform_<endo name>_vs_shocks_*.fig}: shows bar charts +of the sensitivity indices for the ten most important parameters driving +the reduced form coefficients of the selected endogenous variables +(@code{namendo}) versus exogenous variables (@code{namexo}); suffix @code{log} +indicates the results for log-transformed entries; + +@item +@code{<mod_file>_redform_GSA(_log).fig}: shows bar chart of all sensitivity +indices for each parameter: this allows one to notice parameters that +have a minor effect for any of the reduced form coefficients. +@end itemize + +Detailed results of the analyses are shown in the subfolder @code{<mod_file>/GSA/redform_prior}, +where the detailed results of the estimation of the single functional relationships +between parameters @math{\theta} and reduced form coefficient (denoted as @math{y} hereafter) are stored in separate directories +named as: + +@itemize +@item +@code{<namendo>_vs_<namlagendo>}: for the entries of the transition matrix; + +@item +@code{<namendo>_vs_<namexo>}: for entries of the matrix of the shocks. +@end itemize + +The following files are stored in each directory: +@itemize +@item +@code{<mod_file>_<namendo>_vs_<namexo>.fig}: histogram and CDF plot of the MC sample of the individual entry +of the shock matrix, in sample and out of sample fit of the ANOVA model; + +@item +@code{<mod_file>_<namendo>_vs_<namexo>_map_SE.fig}: for entries of the shock matrix it shows graphs of the estimated first order ANOVA terms @math{y = f(\theta_i)} for each deep parameter \math{\theta_i}; + +@item +@code{<mod_file>_<namendo>_vs_<namlagendo>.fig}: histogram and CDF plot of the MC sample of the individual entry +of the transition matrix, in sample and out of sample fit of the ANOVA model; + +@item +@code{<mod_file>_<namendo>_vs_<namlagendo>_map_SE.fig}: for entries of the transition matrix it shows graphs of the estimated first order ANOVA terms @math{y = f(\theta_i)} for each deep parameter @math{\theta_i}; + +@item +@code{<mod_file>_<namendo>_vs_<namexo>_map.mat}, @code{<mod_file>_<namendo>_vs_<namlagendo>_map.mat}: these files store info in the estimation; + +@end itemize + +When option @code{logtrans_redform} is set, the ANOVA estimation is performed using a log-transformation of each @code{y}. +The ANOVA mapping is then transformed back onto the original scale, to allow comparability with the baseline estimation. +Graphs for this log-transformed case, are stored in the same folder in files denoted with the @code{_log} suffix. + +When the option @code{threshold_redform} is set, the analysis is performed via Monte Carlo filtering, by displaying parameters that drive the individual entry @code{y} inside the range specified in @code{threshold_redform}. If no entry is found (or all entries are in the range), the MCF algorithm ignores the range specified in @code{threshold_redform} and performs the analysis splitting the MC sample of @code{y} into deciles. Setting @code{threshold_redform=[-inf inf]} triggers this approach for all @code{y}'s. + +Results are stored in subdirectories of @code{<mod_file>/GSA/redform_prior} named +@itemize +@item +@code{<namendo>_vs_<namlagendo>_threshold}: for the entries of the transition matrix; + +@item +@code{<namendo>_vs_<namexo>_threshold}: for entries of the matrix of the shocks. +@end itemize + +The files saved are named +@itemize +@item +@code{<mod_file>_<namendo>_vs_<namexo>_threshold.fig},@code{<mod_file>_<namendo>_vs_<namlagendo>_threshold.fig}: graphical outputs; +@item +@code{<mod_file>_<namendo>_vs_<namexo>_threshold.mat},@code{<mod_file>_<namendo>_vs_<namlagendo>_threshold.mat}: info on the analysis; + +@end itemize + +@node RMSE +@subsubsection RMSE + +The RMSE analysis can be performed with different types of sampling options: +@enumerate +@item +When @code{pprior=1} and @code{ppost=0}, the toolbox analyzes the RMSEs for +the Monte-Carlo sample obtained by sampling parameters from their prior distributions +(or prior ranges): this analysis provides some hints about +what parameter drives the fit of which observed series, prior to the full +estimation; + +@item +When @code{pprior=0} and @code{ppost=0}, the toolbox analyzes the RMSEs for +a multivariate normal Monte-Carlo sample, with covariance matrix based on +the inverse Hessian at the optimum: this analysis is useful when maximum likelihood +estimation is done (@i{i.e.} no Bayesian estimation); + +@item +When @code{ppost=1} the toolbox analyzes the RMSEs for the posterior sample +obtained by Dynare's Metropolis procedure. +@end enumerate + +The use of cases 2 and 3 requires an estimation step beforehand. To +facilitate the sensitivity analysis after estimation, the @code{dynare_sensitivity} +command also allows you to indicate some options of the @code{estimation} +command. These are: +@itemize @bullet +@item @code{datafile} +@item @code{nobs} +@item @code{first_obs} +@item @code{prefilter} +@item @code{presample} +@item @code{nograph} +@item @code{nodisplay} +@item @code{graph_format} +@item @code{conf_sig} +@item @code{loglinear} +@item @code{mode_file} +@end itemize + +Binary files produced my RMSE analysis are: +@itemize +@item +@code{<mod_file>_prior_*.mat}: these files store the filtered and smoothed + variables for the prior Monte-Carlo sample, generated when doing RMSE analysis + (@code{pprior=1} and @code{ppost=0}); +@item +@code{<mode_file>_mc_*.mat}: these files store the filtered and smoothed variables + for the multivariate normal Monte-Carlo sample, generated when doing + RMSE analysis (@code{pprior=0} and @code{ppost=0}). +@end itemize + +Figure files @code{<mod_file>_rmse_*.fig} store results for the RMSE analysis. + +@itemize +@item +@code{<mod_file>_rmse_prior*.fig}: save results for the analysis using prior +Monte-Carlo samples; + +@item +@code{<mod_file>_rmse_mc*.fig}: save results for the analysis using multivariate +normal Monte-Carlo samples; + +@item +@code{<mod_file>_rmse_post*.fig}: save results for the analysis using Metropolis +posterior samples. +@end itemize + +The following types of figures are saved (we show prior sample to fix ideas, +but the same conventions are used for multivariate normal and posterior): + +@itemize +@item +@code{<mod_file>_rmse_prior_params_*.fig}: for each parameter, plots the cdfs +corresponding to the best 10% RMSEs of each observed series (only those cdfs below the significance threshold @code{alpha_rmse}); + +@item +@code{<mod_file>_rmse_prior_<var_obs>_*.fig}: if a parameter significantly affects the fit of @code{var_obs}, all possible trade-off's with other observables for same parameter are plotted; + +@item +@code{<mod_file>_rmse_prior_<var_obs>_map.fig}: plots the MCF analysis of parameters significantly driving the fit the observed series @code{var_obs}; + +@item +@code{<mod_file>_rmse_prior_lnlik*.fig}: for each observed series, plots +in BLUE the cdf of the log-likelihood corresponding to the best 10% +RMSEs, in RED the cdf of the rest of the sample and in BLACK the +cdf of the full sample; this allows one to see the presence of some +idiosyncratic behavior; + +@item +@code{<mod_file>_rmse_prior_lnpost*.fig}: for each observed series, plots +in BLUE the cdf of the log-posterior corresponding to the best 10% RMSEs, +in RED the cdf of the rest of the sample and in BLACK the cdf of the full +sample; this allows one to see idiosyncratic behavior; + +@item +@code{<mod_file>_rmse_prior_lnprior*.fig}: for each observed series, plots +in BLUE the cdf of the log-prior corresponding to the best 10% RMSEs, +in RED the cdf of the rest of the sample and in BLACK the cdf of the full +sample; this allows one to see idiosyncratic behavior; + +@item +@code{<mod_file>_rmse_prior_lik.fig}: when @code{lik_only=1}, this shows +the MCF tests for the filtering of the best 10% log-likelihood values; + +@item +@code{<mod_file>_rmse_prior_post.fig}: when @code{lik_only=1}, this shows +the MCF tests for the filtering of the best 10% log-posterior values. +@end itemize + +@node Screening Analysis +@subsubsection Screening Analysis + +Screening analysis does not require any additional options with respect to +those listed in @ref{Sampling Options}. The toolbox performs all the +analyses required and displays results. + +The results of the screening analysis with Morris sampling design are stored +in the subfolder @code{<mod_file>/GSA/SCREEN}. The data file @code{<mod_file>_prior} stores +all the information of the analysis (Morris sample, reduced form coefficients, +etc.). + +Screening analysis merely concerns reduced form coefficients. Similar +synthetic bar charts as for the reduced form analysis with Monte-Carlo samples are +saved: +@itemize +@item +@code{<mod_file>_redform_<endo name>_vs_lags_*.fig}: shows bar charts +of the elementary effect tests for the ten most important parameters +driving the reduced form coefficients of the selected endogenous variables +(@code{namendo}) versus lagged endogenous variables (@code{namlagendo}); + +@item +@code{<mod_file>_redform_<endo name>_vs_shocks_*.fig}: shows bar charts +of the elementary effect tests for the ten most important parameters +driving the reduced form coefficients of the selected endogenous variables +(@code{namendo}) versus exogenous variables (@code{namexo}); + +@item +@code{<mod_file>_redform_screen.fig}: shows bar chart of all elementary +effect tests for each parameter: this allows one to identify parameters that +have a minor effect for any of the reduced form coefficients. +@end itemize + +@node Identification Analysis +@subsubsection Identification Analysis + +Setting the option @code{identification=1}, an identification analysis based on +theoretical moments is performed. Sensitivity plots are provided that allow +to infer which parameters are most likely to be less identifiable. + +Prerequisite for properly running all the identification routines, is the keyword +@code{identification}; in the Dynare model file. This keyword triggers +the computation of analytic derivatives of the model with respect to estimated +parameters and shocks. This is required for option @code{morris=2}, +which implements @cite{Iskrev (2010)} identification analysis. + +For example, the placing @code{identification; dynare_sensitivity(identification=1, morris=2);} +in the Dynare model file trigger identification analysis using analytic derivatives +@cite{Iskrev (2010)}, jointly with the mapping of the acceptable region. + +The identification analysis with derivatives can also be triggered by the +commands @code{identification;} This does not do the mapping of +acceptable regions for the model and uses the standard random sampler of Dynare. +It completely offsets any use of the sensitivity analysis toolbox. + + @node Markov-switching SBVAR @section Markov-switching SBVAR -- GitLab