/usr/share/doc/dynare/dynare.html/Sensitivity-and-identification-analysis.html

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<html>
<!-- Copyright (C) 1996-2017, Dynare Team.

Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.3 or
any later version published by the Free Software Foundation; with no
Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts.

A copy of the license can be found at http://www.gnu.org/licenses/fdl.txt. -->
<!-- Created by GNU Texinfo 6.5, http://www.gnu.org/software/texinfo/ -->
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<title>Sensitivity and identification analysis (Dynare Reference Manual)</title>

<meta name="description" content="Sensitivity and identification analysis (Dynare Reference Manual)">
<meta name="keywords" content="Sensitivity and identification analysis (Dynare Reference Manual)">
<meta name="resource-type" content="document">
<meta name="distribution" content="global">
<meta name="Generator" content="texi2any">
<link href="index.html#Top" rel="start" title="Top">
<link href="Command-and-Function-Index.html#Command-and-Function-Index" rel="index" title="Command and Function Index">
<link href="index.html#SEC_Contents" rel="contents" title="Table of Contents">
<link href="The-Model-file.html#The-Model-file" rel="up" title="The Model file">
<link href="Markov_002dswitching-SBVAR.html#Markov_002dswitching-SBVAR" rel="next" title="Markov-switching SBVAR">
<link href="Optimal-policy.html#Optimal-policy" rel="prev" title="Optimal policy">
<style type="text/css">
<!--
a.summary-letter {text-decoration: none}
blockquote.indentedblock {margin-right: 0em}
blockquote.smallindentedblock {margin-right: 0em; font-size: smaller}
blockquote.smallquotation {font-size: smaller}
div.display {margin-left: 3.2em}
div.example {margin-left: 3.2em}
div.lisp {margin-left: 3.2em}
div.smalldisplay {margin-left: 3.2em}
div.smallexample {margin-left: 3.2em}
div.smalllisp {margin-left: 3.2em}
kbd {font-style: oblique}
pre.display {font-family: inherit}
pre.format {font-family: inherit}
pre.menu-comment {font-family: serif}
pre.menu-preformatted {font-family: serif}
pre.smalldisplay {font-family: inherit; font-size: smaller}
pre.smallexample {font-size: smaller}
pre.smallformat {font-family: inherit; font-size: smaller}
pre.smalllisp {font-size: smaller}
span.nolinebreak {white-space: nowrap}
span.roman {font-family: initial; font-weight: normal}
span.sansserif {font-family: sans-serif; font-weight: normal}
ul.no-bullet {list-style: none}
-->
</style>


</head>

<body lang="en">
<a name="Sensitivity-and-identification-analysis"></a>
<div class="header">
<p>
Next: <a href="Markov_002dswitching-SBVAR.html#Markov_002dswitching-SBVAR" accesskey="n" rel="next">Markov-switching SBVAR</a>, Previous: <a href="Optimal-policy.html#Optimal-policy" accesskey="p" rel="prev">Optimal policy</a>, Up: <a href="The-Model-file.html#The-Model-file" accesskey="u" rel="up">The Model file</a> &nbsp; [<a href="index.html#SEC_Contents" title="Table of contents" rel="contents">Contents</a>][<a href="Command-and-Function-Index.html#Command-and-Function-Index" title="Index" rel="index">Index</a>]</p>
</div>
<a name="Sensitivity-and-identification-analysis-1"></a>
<h3 class="section">4.20 Sensitivity and identification analysis</h3>

<p>Dynare provides an interface to the global sensitivity analysis (GSA)
toolbox (developed by the Joint Research Center (JRC) of the European
Commission), which is now part of the official Dynare distribution. The
GSA toolbox can be used to answer the following questions:
</p>
<ol>
<li> What is the domain of structural coefficients assuring the stability and determinacy
of a DSGE model?

</li><li> Which parameters mostly drive the fit of, <i>e.g.</i>, GDP and which the fit of inflation?
Is there any conflict between the optimal fit of one observed series versus another?

</li><li> How to represent in a direct, albeit approximated, form the relationship between
structural parameters and the reduced form of a rational expectations model?
</li></ol>

<p>The discussion of the methodologies and their application is described in
<cite>Ratto (2008)</cite>.
</p>
<p>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.
</p>

<table class="menu" border="0" cellspacing="0">
<tr><td align="left" valign="top">&bull; <a href="#Performing-sensitivity-analysis" accesskey="1">Performing sensitivity analysis</a>:</td><td>&nbsp;&nbsp;</td><td align="left" valign="top">
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#IRF_002fMoment-calibration" accesskey="2">IRF/Moment calibration</a>:</td><td>&nbsp;&nbsp;</td><td align="left" valign="top">
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Performing-identification-analysis" accesskey="3">Performing identification analysis</a>:</td><td>&nbsp;&nbsp;</td><td align="left" valign="top">
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Types-of-analysis-and-output-files" accesskey="4">Types of analysis and output files</a>:</td><td>&nbsp;&nbsp;</td><td align="left" valign="top">
</td></tr>
</table>


<hr>
<a name="Performing-sensitivity-analysis"></a>
<div class="header">
<p>
Next: <a href="#IRF_002fMoment-calibration" accesskey="n" rel="next">IRF/Moment calibration</a>, Up: <a href="#Sensitivity-and-identification-analysis" accesskey="u" rel="up">Sensitivity and identification analysis</a> &nbsp; [<a href="index.html#SEC_Contents" title="Table of contents" rel="contents">Contents</a>][<a href="Command-and-Function-Index.html#Command-and-Function-Index" title="Index" rel="index">Index</a>]</p>
</div>
<a name="Performing-sensitivity-analysis-1"></a>
<h4 class="subsection">4.20.1 Performing sensitivity analysis</h4>

<a name="dynare_005fsensitivity"></a><dl>
<dt><a name="index-dynare_005fsensitivity"></a>Command: <strong>dynare_sensitivity</strong> <em>;</em></dt>
<dt><a name="index-dynare_005fsensitivity-1"></a>Command: <strong>dynare_sensitivity</strong> <em>(<var>OPTIONS</var>&hellip;);</em></dt>
<dd>
<p><em>Description</em>
</p>
<p>This command triggers sensitivity analysis on a DSGE model.
</p>
<p><em>Options</em>
<em>Sampling Options</em>
<a name="Sampling-Options"></a></p><dl compact="compact">
<dt><code>Nsam = <var>INTEGER</var></code></dt>
<dd><p>Size of the Monte-Carlo sample. Default: <code>2048</code>
</p>
</dd>
<dt><code>ilptau = <var>INTEGER</var></code></dt>
<dd><p>If equal to <code>1</code>, use <SPAN CLASS="MATH"><IMG
 WIDTH="34" HEIGHT="30" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_125.png"
 ALT="$LP_\tau$"></SPAN> quasi-Monte-Carlo.
If equal to <code>0</code>, use LHS Monte-Carlo. Default: <code>1</code>
</p>
</dd>
<dt><code>pprior = <var>INTEGER</var></code></dt>
<dd><p>If equal to <code>1</code>, sample from the prior distributions.
If equal to <code>0</code>, sample from the multivariate normal <!-- MATH
 $N(\bar{\theta},\Sigma)$
 -->
<SPAN CLASS="MATH"><IMG
 WIDTH="58" HEIGHT="36" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_126.png"
 ALT="$N(\bar{\theta},\Sigma)$"></SPAN>,
where <SPAN CLASS="MATH"><IMG
 WIDTH="12" HEIGHT="17" ALIGN="BOTTOM" BORDER="0"
 SRC="dynare.html_127.png"
 ALT="$\bar{\theta}$"></SPAN> is the posterior mode and <SPAN CLASS="MATH"><IMG
 WIDTH="69" HEIGHT="16" ALIGN="BOTTOM" BORDER="0"
 SRC="dynare.html_128.png"
 ALT="$\Sigma=H^{-1}$"></SPAN>, <SPAN CLASS="MATH"><IMG
 WIDTH="19" HEIGHT="14" ALIGN="BOTTOM" BORDER="0"
 SRC="dynare.html_129.png"
 ALT="$H$"></SPAN>
is the Hessian at the mode. Default: <code>1</code>
</p>
</dd>
<dt><code>prior_range = <var>INTEGER</var></code></dt>
<dd><p>If equal to <code>1</code>, sample uniformly from prior ranges.
If equal to <code>0</code>, sample from prior distributions. Default: <code>1</code>
</p>
</dd>
<dt><code>morris = <var>INTEGER</var></code></dt>
<dd><a name="morris"></a><p>If equal to <code>0</code>, ANOVA mapping (Type I error)
If equal to <code>1</code>, Screening analysis (Type II error)
If equal to <code>2</code>, Analytic derivatives (similar to Type II error, only valid when
<code>identification=1</code>).Default: <code>1</code> when <code>identification=1</code>, <code>0</code> otherwise
</p>
</dd>
<dt><code>morris_nliv = <var>INTEGER</var></code></dt>
<dd><a name="morris_005fnliv"></a><p>Number of levels in Morris design. Default: <code>6</code>
</p>
</dd>
<dt><code>morris_ntra = <var>INTEGER</var></code></dt>
<dd><a name="morris_005fntra"></a><p>Number trajectories in Morris design. Default: <code>20</code>
</p>
</dd>
<dt><code>ppost = <var>INTEGER</var></code></dt>
<dd><p>If equal to <code>1</code>, use Metropolis posterior sample.
If equal to <code>0</code>, do not use Metropolis posterior sample. NB: This
overrides any other sampling option. Default: <code>0</code>
</p>
</dd>
<dt><code>neighborhood_width = <var>DOUBLE</var></code></dt>
<dd><p>When <code>pprior=0</code> and <code>ppost=0</code>, allows for the sampling of
parameters around the value specified in the <code>mode_file</code>, in the range
<code>xparam1</code><!-- MATH
 $\pm\left|@code{xparam1}\times@code{neighborhood_width}\right|$
 -->
<SPAN CLASS="MATH"><IMG
 WIDTH="330" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_130.png"
 ALT="$\pm\left\vert@code{xparam1}\times@code{neighborhood_width}\right\vert$"></SPAN>. Default: <code>0</code>
</p>
</dd>
</dl>
<p><em>Stability Mapping Options</em>
</p><dl compact="compact">
<dt><code>stab = <var>INTEGER</var></code></dt>
<dd><p>If equal to <code>1</code>, perform stability mapping.
If equal to <code>0</code>, do not perform stability mapping. Default: <code>1</code>
</p>
</dd>
<dt><code>load_stab = <var>INTEGER</var></code></dt>
<dd><p>If equal to <code>1</code>, load a previously created sample.
If equal to <code>0</code>, generate a new sample. Default: <code>0</code>
</p>
</dd>
<dt><code>alpha2_stab = <var>DOUBLE</var></code></dt>
<dd><p>Critical value for correlations <SPAN CLASS="MATH"><IMG
 WIDTH="13" HEIGHT="30" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_131.png"
 ALT="$\rho$"></SPAN> in filtered samples:
plot couples of parmaters with <!-- MATH
 $\left|\rho\right|>$
 -->
<SPAN CLASS="MATH"><IMG
 WIDTH="38" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_132.png"
 ALT="$\left\vert\rho\right\vert&gt;$"></SPAN> <code>alpha2_stab</code>.
Default: <code>0</code>
</p>
</dd>
<dt><code>pvalue_ks = <var>DOUBLE</var></code></dt>
<dd><p>The threshold <SPAN CLASS="MATH"><IMG
 WIDTH="51" HEIGHT="30" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_133.png"
 ALT="$pvalue$"></SPAN> for significant Kolmogorov-Smirnov test (<i>i.e.</i> plot parameters with
<SPAN CLASS="MATH"><IMG
 WIDTH="68" HEIGHT="30" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_134.png"
 ALT="$pvalue&lt;$"></SPAN> <code>pvalue_ks</code>). Default: <code>0.001</code>
</p>
</dd>
<dt><code>pvalue_corr = <var>DOUBLE</var></code></dt>
<dd><p>The threshold <SPAN CLASS="MATH"><IMG
 WIDTH="51" HEIGHT="30" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_133.png"
 ALT="$pvalue$"></SPAN> for significant correlation in filtered samples
(<i>i.e.</i> plot bivariate samples when <SPAN CLASS="MATH"><IMG
 WIDTH="68" HEIGHT="30" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_134.png"
 ALT="$pvalue&lt;$"></SPAN> <code>pvalue_corr</code>). Default: <code>1e-5</code>
</p>
</dd>
</dl>
<p><em>Reduced Form Mapping Options</em>
</p><dl compact="compact">
<dt><code>redform = <var>INTEGER</var></code></dt>
<dd><p>If equal to <code>1</code>, prepare Monte-Carlo sample of reduced form matrices.
If equal to <code>0</code>, do not prepare Monte-Carlo sample of reduced form matrices. Default: <code>0</code>
</p>
</dd>
<dt><code>load_redform = <var>INTEGER</var></code></dt>
<dd><p>If equal to <code>1</code>, load previously estimated mapping.
If equal to <code>0</code>, estimate the mapping of the reduced form model. Default: <code>0</code>
</p>
</dd>
<dt><code>logtrans_redform = <var>INTEGER</var></code></dt>
<dd><p>If equal to <code>1</code>, use log-transformed entries.
If equal to <code>0</code>, use raw entries. Default: <code>0</code>
</p>
</dd>
<dt><code>threshold_redform = [<var>DOUBLE</var> <var>DOUBLE</var>]</code></dt>
<dd><p>The range over which the filtered Monte-Carlo entries of the reduced form coefficients
should be analyzed. The first number is the lower bound and the second is the upper bound.
An empty vector indicates that these entries will not be filtered. Default: <code>empty</code>
</p>
</dd>
<dt><code>ksstat_redform = <var>DOUBLE</var></code></dt>
<dd><p>Critical value for Smirnov statistics <SPAN CLASS="MATH"><IMG
 WIDTH="13" HEIGHT="15" ALIGN="BOTTOM" BORDER="0"
 SRC="dynare.html_135.png"
 ALT="$d$"></SPAN> when reduced form entries
are filtered. Default: <code>0.001</code>
</p>
</dd>
<dt><code>alpha2_redform = <var>DOUBLE</var></code></dt>
<dd><p>Critical value for correlations <SPAN CLASS="MATH"><IMG
 WIDTH="13" HEIGHT="30" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_131.png"
 ALT="$\rho$"></SPAN>  when reduced form entries
are filtered. Default: <code>1e-5</code>
</p>
</dd>
<dt><code>namendo = (<var>VARIABLE_NAME</var>&hellip;)</code></dt>
<dd><p>List of endogenous variables. &lsquo;<code>:</code>&rsquo; indicates all endogenous variables.
Default: <code>empty</code>
</p>
</dd>
<dt><code>namlagendo = (<var>VARIABLE_NAME</var>&hellip;)</code></dt>
<dd><p>List of lagged endogenous variables. &lsquo;<code>:</code>&rsquo; indicates all lagged endogenous variables.
Analyze entries <code>[namendo</code><SPAN CLASS="MATH"><IMG
 WIDTH="17" HEIGHT="30" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_136.png"
 ALT="$\times$"></SPAN><code>namlagendo]</code> Default: <code>empty</code>
</p>
</dd>
<dt><code>namexo = (<var>VARIABLE_NAME</var>&hellip;)</code></dt>
<dd><p>List of exogenous variables. &lsquo;<code>:</code>&rsquo; indicates all exogenous variables.
Analyze entries <code>[namendo</code><SPAN CLASS="MATH"><IMG
 WIDTH="17" HEIGHT="30" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_136.png"
 ALT="$\times$"></SPAN><code>namexo]</code>. Default: <code>empty</code>
</p>
</dd>
</dl>
<p><em>RMSE Options</em>
</p><dl compact="compact">
<dt><code>rmse = <var>INTEGER</var></code></dt>
<dd><p>If equal to <code>1</code>, perform RMSE analysis.
If equal to <code>0</code>, do not perform RMSE analysis. Default: <code>0</code>
</p>
</dd>
<dt><code>load_rmse = <var>INTEGER</var></code></dt>
<dd><p>If equal to <code>1</code>, load previous RMSE analysis.
If equal to <code>0</code>, make a new RMSE analysis. Default: <code>0</code>
</p>
</dd>
<dt><code>lik_only = <var>INTEGER</var></code></dt>
<dd><p>If equal to <code>1</code>, compute only likelihood and posterior.
If equal to <code>0</code>, compute RMSE&rsquo;s for all observed series. Default: <code>0</code>
</p>
</dd>
<dt><code>var_rmse = (<var>VARIABLE_NAME</var>&hellip;)</code></dt>
<dd><p>List of observed series to be considered. &lsquo;<code>:</code>&rsquo; indicates all observed
variables. Default: <code>varobs</code>
</p>
</dd>
<dt><code>pfilt_rmse = <var>DOUBLE</var></code></dt>
<dd><p>Filtering threshold for RMSE&rsquo;s. Default: <code>0.1</code>
</p>
</dd>
<dt><code>istart_rmse = <var>INTEGER</var></code></dt>
<dd><p>Value at which to start computing RMSE&rsquo;s (use <code>2</code> to avoid big intitial
error). Default: <code>presample+1</code>
</p>
</dd>
<dt><code>alpha_rmse = <var>DOUBLE</var></code></dt>
<dd><p>Critical value for Smirnov statistics <SPAN CLASS="MATH"><IMG
 WIDTH="13" HEIGHT="15" ALIGN="BOTTOM" BORDER="0"
 SRC="dynare.html_135.png"
 ALT="$d$"></SPAN>: plot parameters with
<SPAN CLASS="MATH"><IMG
 WIDTH="30" HEIGHT="30" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_137.png"
 ALT="$d&gt;$"></SPAN> <code>alpha_rmse</code>. Default: <code>0.001</code>
</p>
</dd>
<dt><code>alpha2_rmse = <var>DOUBLE</var></code></dt>
<dd><p>Critical value for correlation <SPAN CLASS="MATH"><IMG
 WIDTH="13" HEIGHT="30" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_131.png"
 ALT="$\rho$"></SPAN>: plot couples of parmaters with
<!-- MATH
 $\left|\rho\right|=$
 -->
<SPAN CLASS="MATH"><IMG
 WIDTH="38" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_138.png"
 ALT="$\left\vert\rho\right\vert=$"></SPAN> <code>alpha2_rmse</code>. Default: <code>1e-5</code>
</p>
</dd>
<dt><code>datafile = <var>FILENAME</var></code></dt>
<dd><p>See <a href="Estimation.html#datafile">datafile</a>.
</p>
</dd>
<dt><code>nobs = <var>INTEGER</var></code></dt>
<dt><code>nobs = [<var>INTEGER1</var>:<var>INTEGER2</var>]</code></dt>
<dd><p>See <a href="Estimation.html#nobs">nobs</a>.
</p>
</dd>
<dt><code>first_obs = <var>INTEGER</var></code></dt>
<dd><p>See <a href="Estimation.html#first_005fobs">first_obs</a>.
</p>
</dd>
<dt><code>prefilter = <var>INTEGER</var></code></dt>
<dd><p>See <a href="Estimation.html#prefilter">prefilter</a>.
</p>
</dd>
<dt><code>presample = <var>INTEGER</var></code></dt>
<dd><p>See <a href="Estimation.html#presample">presample</a>.
</p>
</dd>
<dt><code>nograph</code></dt>
<dd><p>See <a href="Stochastic-solution-and-simulation.html#nograph">nograph</a>.
</p>
</dd>
<dt><code>nodisplay</code></dt>
<dd><p>See <a href="Stochastic-solution-and-simulation.html#nodisplay">nodisplay</a>.
</p>
</dd>
<dt><code>graph_format = <var>FORMAT</var></code></dt>
<dt><code>graph_format = ( <var>FORMAT</var>, <var>FORMAT</var>&hellip; )</code></dt>
<dd><p>See <a href="Stochastic-solution-and-simulation.html#graph_005fformat">graph_format</a>.
</p>
</dd>
<dt><code>conf_sig = <var>DOUBLE</var></code></dt>
<dd><p>See <a href="Forecasting.html#conf_005fsig">conf_sig</a>.
</p>
</dd>
<dt><code>loglinear</code></dt>
<dd><p>See <a href="Estimation.html#loglinear">loglinear</a>.
</p>
</dd>
<dt><code>mode_file = <var>FILENAME</var></code></dt>
<dd><p>See <a href="Estimation.html#mode_005ffile">mode_file</a>.
</p>
</dd>
<dt><code>kalman_algo = <var>INTEGER</var></code></dt>
<dd><p>See <a href="Estimation.html#kalman_005falgo">kalman_algo</a>.
</p>
</dd>
</dl>
<p><em>Identification Analysis Options</em>
</p><dl compact="compact">
<dt><code>identification = <var>INTEGER</var></code></dt>
<dd><p>If equal to <code>1</code>, performs identification anlysis (forcing <code>redform=0</code> and <code>morris=1</code>)
If equal to <code>0</code>, no identification analysis. Default: <code>0</code>
</p>
</dd>
<dt><code>morris = <var>INTEGER</var></code></dt>
<dd><p>See <a href="#morris">morris</a>.
</p>
</dd>
<dt><code>morris_nliv = <var>INTEGER</var></code></dt>
<dd><p>See <a href="#morris_005fnliv">morris_nliv</a>.
</p>
</dd>
<dt><code>morris_ntra = <var>INTEGER</var></code></dt>
<dd><p>See <a href="#morris_005fntra">morris_ntra</a>.
</p>
</dd>
<dt><code>load_ident_files = <var>INTEGER</var></code></dt>
<dd><p>Loads previously performed identification analysis. Default: <code>0</code>
</p>
</dd>
<dt><code>useautocorr = <var>INTEGER</var></code></dt>
<dd><p>Use autocorrelation matrices in place of autocovariance matrices in moments
for identification analysis. Default: <code>0</code>
</p>
</dd>
<dt><code>ar = <var>INTEGER</var></code></dt>
<dd><p>Maximum number of lags for moments in identification analysis. Default: <code>1</code>
</p>
</dd>
<dt><code>diffuse_filter = <var>INTEGER</var></code></dt>
<dd><p>See <a href="Estimation.html#diffuse_005ffilter">diffuse_filter</a>.
</p>
</dd>
</dl>

</dd></dl>

<hr>
<a name="IRF_002fMoment-calibration"></a>
<div class="header">
<p>
Next: <a href="#Performing-identification-analysis" accesskey="n" rel="next">Performing identification analysis</a>, Previous: <a href="#Performing-sensitivity-analysis" accesskey="p" rel="prev">Performing sensitivity analysis</a>, Up: <a href="#Sensitivity-and-identification-analysis" accesskey="u" rel="up">Sensitivity and identification analysis</a> &nbsp; [<a href="index.html#SEC_Contents" title="Table of contents" rel="contents">Contents</a>][<a href="Command-and-Function-Index.html#Command-and-Function-Index" title="Index" rel="index">Index</a>]</p>
</div>
<a name="IRF_002fMoment-calibration-1"></a>
<h4 class="subsection">4.20.2 IRF/Moment calibration</h4>

<p>The <code>irf_calibration</code> and <code>moment_calibration</code> blocks allow imposing implicit &ldquo;endogenous&rdquo; priors 
about IRFs and moments on the model. The way it works internally is that 
any parameter draw that is inconsistent with the &ldquo;calibration&rdquo; provided in these blocks is discarded, <i>i.e.</i> assigned a prior density of <SPAN CLASS="MATH"><IMG
 WIDTH="12" HEIGHT="14" ALIGN="BOTTOM" BORDER="0"
 SRC="dynare.html_2.png"
 ALT="$0$"></SPAN>. 
In the context of <code>dynare_sensitivity</code>, these restrictions allow tracing out which parameters are driving the model to
satisfy or violate the given restrictions.
</p>
<p>IRF and moment calibration can be defined in <code>irf_calibration</code> and <code>moment_calibration</code> blocks:
</p>
<dl>
<dt><a name="index-irf_005fcalibration"></a>Block: <strong>irf_calibration</strong> <em>;</em></dt>
<dt><a name="index-irf_005fcalibration-1"></a>Block: <strong>irf_calibration</strong> <em>(<var>OPTIONS</var>&hellip;);</em></dt>
<dd>
<p><em>Description</em>
</p>
<p>This block allows defining IRF calibration criteria and is terminated by <code>end;</code>. 
To set IRF sign restrictions, the following syntax is used
</p><div class="example">
<pre class="example"><var>VARIABLE_NAME</var>(<var>INTEGER</var>),<var>EXOGENOUS_NAME</var>, -;
<var>VARIABLE_NAME</var>(<var>INTEGER</var>:<var>INTEGER</var>),<var>EXOGENOUS_NAME</var>, +;
</pre></div>
<p>To set IRF restrictions with specific intervals, the following syntax is used
</p><div class="example">
<pre class="example"><var>VARIABLE_NAME</var>(<var>INTEGER</var>),<var>EXOGENOUS_NAME</var>, [<var>DOUBLE</var> <var>DOUBLE</var>];
<var>VARIABLE_NAME</var>(<var>INTEGER</var>:<var>INTEGER</var>),<var>EXOGENOUS_NAME</var>, [<var>DOUBLE</var> <var>DOUBLE</var>];
</pre></div>

<p>When <code>(<var>INTEGER</var>:<var>INTEGER</var>)</code> is used, the restriction is considered to be fulfilled by a logical OR.
A list of restrictions must always be fulfilled with logical AND.
</p>
<p><em>Options</em>
</p>
<dl compact="compact">
<dt><code>relative_irf</code></dt>
<dd><p>See <a href="Stochastic-solution-and-simulation.html#relative_005firf">relative_irf</a>.
</p>
</dd>
</dl>

<p><em>Example</em>
</p>
<div class="example">
<pre class="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;
</pre></div>

</dd></dl>

<dl>
<dt><a name="index-moment_005fcalibration"></a>Block: <strong>moment_calibration</strong> <em>;</em></dt>
<dt><a name="index-moment_005fcalibration-1"></a>Block: <strong>moment_calibration</strong> <em>(<var>OPTIONS</var>&hellip;);</em></dt>
<dd>
<p><em>Description</em>
</p>
<p>This block allows defining moment calibration criteria. This block is terminated by <code>end;</code>, and contains lines of the
form:
</p><div class="example">
<pre class="example"><var>VARIABLE_NAME1</var>,<var>VARIABLE_NAME2</var>(+/-<var>INTEGER</var>), [<var>DOUBLE</var> <var>DOUBLE</var>];
<var>VARIABLE_NAME1</var>,<var>VARIABLE_NAME2</var>(+/-<var>INTEGER</var>), +/-;
<var>VARIABLE_NAME1</var>,<var>VARIABLE_NAME2</var>(+/-(<var>INTEGER</var>:<var>INTEGER</var>)), [<var>DOUBLE</var> <var>DOUBLE</var>];
<var>VARIABLE_NAME1</var>,<var>VARIABLE_NAME2</var>((-<var>INTEGER</var>:+<var>INTEGER</var>)), [<var>DOUBLE</var> <var>DOUBLE</var>];
</pre></div>

<p>When <code>(<var>INTEGER</var>:<var>INTEGER</var>)</code> is used, the restriction is considered to be fulfilled by a logical OR.
A list of restrictions must always be fulfilled with logical AND.
</p>
<p><em>Example</em>
</p>
<div class="example">
<pre class="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;
</pre></div>

</dd></dl>
<hr>
<a name="Performing-identification-analysis"></a>
<div class="header">
<p>
Next: <a href="#Types-of-analysis-and-output-files" accesskey="n" rel="next">Types of analysis and output files</a>, Previous: <a href="#IRF_002fMoment-calibration" accesskey="p" rel="prev">IRF/Moment calibration</a>, Up: <a href="#Sensitivity-and-identification-analysis" accesskey="u" rel="up">Sensitivity and identification analysis</a> &nbsp; [<a href="index.html#SEC_Contents" title="Table of contents" rel="contents">Contents</a>][<a href="Command-and-Function-Index.html#Command-and-Function-Index" title="Index" rel="index">Index</a>]</p>
</div>
<a name="Performing-identification-analysis-1"></a>
<h4 class="subsection">4.20.3 Performing identification analysis</h4>

<a name="identification"></a><dl>
<dt><a name="index-identification"></a>Command: <strong>identification</strong> <em>;</em></dt>
<dt><a name="index-identification-1"></a>Command: <strong>identification</strong> <em>(<var>OPTIONS</var>&hellip;);</em></dt>
<dd>
<p><em>Description</em>
</p>
<p>This command triggers identification analysis.
</p>
<p><em>Options</em>
</p>
<dl compact="compact">
<dt><code>ar = <var>INTEGER</var></code></dt>
<dd><p>Number of lags of computed autocorrelations (theoretical moments). Default: <code>1</code>
</p>
</dd>
<dt><code>useautocorr = <var>INTEGER</var></code></dt>
<dd><p>If equal to <code>1</code>, compute derivatives of autocorrelation. If equal
to <code>0</code>, compute derivatives of autocovariances. Default: <code>0</code>
</p>
</dd>
<dt><code>load_ident_files = <var>INTEGER</var></code></dt>
<dd><p>If equal to <code>1</code>, allow Dynare to load previously
computed analyzes. Default: <code>0</code>
</p>
</dd>
<dt><code>prior_mc = <var>INTEGER</var></code></dt>
<dd><p>Size of Monte-Carlo sample. Default: <code>1</code>
</p>
</dd>
<dt><code>prior_range = <var>INTEGER</var></code></dt>
<dd><p>Triggers uniform sample within the range implied by the prior specifications (when
<code>prior_mc&gt;1</code>). Default: <code>0</code>
</p>
</dd>
<dt><code>advanced = <var>INTEGER</var></code></dt>
<dd><p>Shows a more detailed analysis, comprised of an analysis for the linearized rational
expectation model as well as the associated reduced form solution. Further performs a brute
force search of the groups of parameters best reproducing the behavior of each single parameter.
The maximum dimension of the group searched is triggered by <code>max_dim_cova_group</code>. Default: <code>0</code>
</p>
</dd>
<dt><code>max_dim_cova_group = <var>INTEGER</var></code></dt>
<dd><p>In the brute force search (performed when <code>advanced=1</code>) this option sets the maximum dimension of groups
of parameters that best reproduce the behavior of each single model parameter. Default: <code>2</code>
</p>
</dd>
<dt><code>periods = <var>INTEGER</var></code></dt>
<dd><p>When the analytic Hessian is not available (<i>i.e.</i> with missing values or diffuse
Kalman filter or univariate Kalman filter), this triggers the length of stochastic simulation
to compute Simulated Moments Uncertainty. Default: <code>300</code>
</p>
</dd>
<dt><code>replic = <var>INTEGER</var></code></dt>
<dd><p>When the analytic Hessian is not available, this triggers the number of replicas
to compute Simulated Moments Uncertainty. Default: <code>100</code>
</p>
</dd>
<dt><code>gsa_sample_file = <var>INTEGER</var></code></dt>
<dd><p>If equal to <code>0</code>, do not use sample file.
If equal to <code>1</code>, triggers gsa prior sample.
If equal to <code>2</code>, triggers gsa Monte-Carlo sample (<i>i.e.</i> loads a sample corresponding to
<code>pprior=0</code> and <code>ppost=0</code> in the <code>dynare_sensitivity</code> options). Default: <code>0</code>
</p>
</dd>
<dt><code>gsa_sample_file = <var>FILENAME</var></code></dt>
<dd><p>Uses the provided path to a specific user defined sample file. Default: <code>0</code>
</p>
</dd>
<dt><code>parameter_set = <code>calibration</code> | <code>prior_mode</code> | <code>prior_mean</code> | <code>posterior_mode</code> | <code>posterior_mean</code> | <code>posterior_median</code></code></dt>
<dd><p>Specify the parameter set to use. Default: <code>prior_mean</code>
</p>
</dd>
<dt><code>lik_init = <var>INTEGER</var></code></dt>
<dd><p>See <a href="Estimation.html#lik_005finit">lik_init</a>.
</p>
</dd>
<dt><code>kalman_algo = <var>INTEGER</var></code></dt>
<dd><p>See <a href="Estimation.html#kalman_005falgo">kalman_algo</a>.
</p>
</dd>
<dt><code>nograph</code></dt>
<dd><p>See <a href="Stochastic-solution-and-simulation.html#nograph">nograph</a>.
</p>
</dd>
<dt><code>nodisplay</code></dt>
<dd><p>See <a href="Stochastic-solution-and-simulation.html#nodisplay">nodisplay</a>.
</p>
</dd>
<dt><code>graph_format = <var>FORMAT</var></code></dt>
<dt><code>graph_format = ( <var>FORMAT</var>, <var>FORMAT</var>&hellip; )</code></dt>
<dd><p>See <a href="Stochastic-solution-and-simulation.html#graph_005fformat">graph_format</a>.
</p>
</dd>
</dl>

</dd></dl>

<hr>
<a name="Types-of-analysis-and-output-files"></a>
<div class="header">
<p>
Previous: <a href="#Performing-identification-analysis" accesskey="p" rel="prev">Performing identification analysis</a>, Up: <a href="#Sensitivity-and-identification-analysis" accesskey="u" rel="up">Sensitivity and identification analysis</a> &nbsp; [<a href="index.html#SEC_Contents" title="Table of contents" rel="contents">Contents</a>][<a href="Command-and-Function-Index.html#Command-and-Function-Index" title="Index" rel="index">Index</a>]</p>
</div>
<a name="Types-of-analysis-and-output-files-1"></a>
<h4 class="subsection">4.20.4 Types of analysis and output files</h4>

<p>The sensitivity analysis toolbox includes several types of analyses.
Sensitivity analysis results are saved locally in <code>&lt;mod_file&gt;/gsa</code>,
where <code>&lt;mod_file&gt;.mod</code> is the name of the DYNARE model file.
</p>
<table class="menu" border="0" cellspacing="0">
<tr><td align="left" valign="top">&bull; <a href="#Sampling" accesskey="1">Sampling</a>:</td><td>&nbsp;&nbsp;</td><td align="left" valign="top">
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Stability-Mapping" accesskey="2">Stability Mapping</a>:</td><td>&nbsp;&nbsp;</td><td align="left" valign="top">
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#IRF_002fMoment-restrictions" accesskey="3">IRF/Moment restrictions</a>:</td><td>&nbsp;&nbsp;</td><td align="left" valign="top">
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Reduced-Form-Mapping" accesskey="4">Reduced Form Mapping</a>:</td><td>&nbsp;&nbsp;</td><td align="left" valign="top">
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#RMSE" accesskey="5">RMSE</a>:</td><td>&nbsp;&nbsp;</td><td align="left" valign="top">
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Screening-Analysis" accesskey="6">Screening Analysis</a>:</td><td>&nbsp;&nbsp;</td><td align="left" valign="top">
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Identification-Analysis" accesskey="7">Identification Analysis</a>:</td><td>&nbsp;&nbsp;</td><td align="left" valign="top">
</td></tr>
</table>

<hr>
<a name="Sampling"></a>
<div class="header">
<p>
Next: <a href="#Stability-Mapping" accesskey="n" rel="next">Stability Mapping</a>, Up: <a href="#Types-of-analysis-and-output-files" accesskey="u" rel="up">Types of analysis and output files</a> &nbsp; [<a href="index.html#SEC_Contents" title="Table of contents" rel="contents">Contents</a>][<a href="Command-and-Function-Index.html#Command-and-Function-Index" title="Index" rel="index">Index</a>]</p>
</div>
<a name="Sampling-1"></a>
<h4 class="subsubsection">4.20.4.1 Sampling</h4>

<p>The following binary files are produced:
</p><ul>
<li> <code>&lt;mod_file&gt;_prior.mat</code>: this file stores information about the analyses
performed sampling from the prior, <i>i.e.</i> <code>pprior=1</code> and <code>ppost=0</code>;

</li><li> <code>&lt;mod_file&gt;_mc.mat</code>: this file stores information about the analyses performed
sampling from multivariate normal, <i>i.e.</i> <code>pprior=0</code> and <code>ppost=0</code>;

</li><li> <code>&lt;mod_file&gt;_post.mat</code>: this file stores information about analyses performed
using the Metropolis posterior sample, <i>i.e.</i> <code>ppost=1</code>.
</li></ul>

<hr>
<a name="Stability-Mapping"></a>
<div class="header">
<p>
Next: <a href="#IRF_002fMoment-restrictions" accesskey="n" rel="next">IRF/Moment restrictions</a>, Previous: <a href="#Sampling" accesskey="p" rel="prev">Sampling</a>, Up: <a href="#Types-of-analysis-and-output-files" accesskey="u" rel="up">Types of analysis and output files</a> &nbsp; [<a href="index.html#SEC_Contents" title="Table of contents" rel="contents">Contents</a>][<a href="Command-and-Function-Index.html#Command-and-Function-Index" title="Index" rel="index">Index</a>]</p>
</div>
<a name="Stability-Mapping-1"></a>
<h4 class="subsubsection">4.20.4.2 Stability Mapping</h4>

<p>Figure files produced are of the form <code>&lt;mod_file&gt;_prior_*.fig</code> and store results
for stability mapping from prior Monte-Carlo samples:
</p><ul>
<li> <code>&lt;mod_file&gt;_prior_stable.fig</code>: 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.</i> either
instability or indeterminacy or the solution could not be found (<i>e.g.</i>
the steady state solution could not be found by the solver);

</li><li> <code>&lt;mod_file&gt;_prior_indeterm.fig</code>: 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);

</li><li> <code>&lt;mod_file&gt;_prior_unstable.fig</code>: 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);

</li><li> <code>&lt;mod_file&gt;_prior_wrong.fig</code>: 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.</i>
the steady state solution could not be found by the solver - red color) 
with the cdf of the rest of the sample (blue color);

</li><li> <code>&lt;mod_file&gt;_prior_calib.fig</code>: 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);

</li></ul>

<p>Similar conventions apply for <code>&lt;mod_file&gt;_mc_*.fig</code> files, obtained when
samples from multivariate normal are used.
</p>
<hr>
<a name="IRF_002fMoment-restrictions"></a>
<div class="header">
<p>
Next: <a href="#Reduced-Form-Mapping" accesskey="n" rel="next">Reduced Form Mapping</a>, Previous: <a href="#Stability-Mapping" accesskey="p" rel="prev">Stability Mapping</a>, Up: <a href="#Types-of-analysis-and-output-files" accesskey="u" rel="up">Types of analysis and output files</a> &nbsp; [<a href="index.html#SEC_Contents" title="Table of contents" rel="contents">Contents</a>][<a href="Command-and-Function-Index.html#Command-and-Function-Index" title="Index" rel="index">Index</a>]</p>
</div>
<a name="IRF_002fMoment-restrictions-1"></a>
<h4 class="subsubsection">4.20.4.3 IRF/Moment restrictions</h4>

<p>The following binary files are produced:
</p><ul>
<li> <code>&lt;mod_file&gt;_prior_restrictions.mat</code>: this file stores information about the IRF/moment restriction analysis
performed sampling from the prior ranges, <i>i.e.</i> <code>pprior=1</code> and <code>ppost=0</code>;

</li><li> <code>&lt;mod_file&gt;_mc_restrictions.mat</code>: this file stores information about the IRF/moment restriction analysis performed
sampling from multivariate normal, <i>i.e.</i> <code>pprior=0</code> and <code>ppost=0</code>;

</li><li> <code>&lt;mod_file&gt;_post_restrictions.mat</code>: this file stores information about IRF/moment restriction analysis performed
using the Metropolis posterior sample, <i>i.e.</i> <code>ppost=1</code>.
</li></ul>

<p>Figure files produced are of the form <code>&lt;mod_file&gt;_prior_irf_calib_*.fig</code> and <code>&lt;mod_file&gt;_prior_moment_calib_*.fig</code> and store results
for mapping restrictions from prior Monte-Carlo samples:
</p><ul>
<li> <code>&lt;mod_file&gt;_prior_irf_calib_&lt;ENDO_NAME&gt;_vs_&lt;EXO_NAME&gt;_&lt;PERIOD&gt;.fig</code>: 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>&lt;ENDO_NAME&gt;</code> vs. <code>&lt;EXO_NAME&gt;</code> at period(s) <code>&lt;PERIOD&gt;</code> is matched (blue color)
with the cdf where the IRF restriction is NOT matched (red color)

</li><li> <code>&lt;mod_file&gt;_prior_irf_calib_&lt;ENDO_NAME&gt;_vs_&lt;EXO_NAME&gt;_ALL.fig</code>: 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>&lt;ENDO_NAME&gt;</code> vs. <code>&lt;EXO_NAME&gt;</code> are matched (blue color)
with the cdf where the IRF restriction is NOT matched (red color)

</li><li> <code>&lt;mod_file&gt;_prior_irf_restrictions.fig</code>: plots visual information on the IRF restrictions 
compared to the actual Monte Carlo realization from prior sample.

</li><li> <code>&lt;mod_file&gt;_prior_moment_calib_&lt;ENDO_NAME1&gt;_vs_&lt;ENDO_NAME2&gt;_&lt;LAG&gt;.fig</code>: 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>&lt;ENDO_NAME1&gt;</code> vs. <code>&lt;ENDO_NAME2&gt;</code> at lag(s) <code>&lt;LAG&gt;</code> is matched (blue color)
with the cdf where the IRF restriction is NOT matched (red color)

</li><li> <code>&lt;mod_file&gt;_prior_moment_calib_&lt;ENDO_NAME&gt;_vs_&lt;EXO_NAME&gt;_ALL.fig</code>: 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>&lt;ENDO_NAME1&gt;</code> vs. <code>&lt;ENDO_NAME2&gt;</code> are matched (blue color)
with the cdf where the IRF restriction is NOT matched (red color)

</li><li> <code>&lt;mod_file&gt;_prior_moment_restrictions.fig</code>: plots visual information on the moment restrictions 
compared to the actual Monte Carlo realization from prior sample.

</li></ul>

<p>Similar conventions apply for <code>&lt;mod_file&gt;_mc_*.fig</code> and <code>&lt;mod_file&gt;_post_*.fig</code> files, obtained when
samples from multivariate normal or from posterior are used.
</p>
<hr>
<a name="Reduced-Form-Mapping"></a>
<div class="header">
<p>
Next: <a href="#RMSE" accesskey="n" rel="next">RMSE</a>, Previous: <a href="#IRF_002fMoment-restrictions" accesskey="p" rel="prev">IRF/Moment restrictions</a>, Up: <a href="#Types-of-analysis-and-output-files" accesskey="u" rel="up">Types of analysis and output files</a> &nbsp; [<a href="index.html#SEC_Contents" title="Table of contents" rel="contents">Contents</a>][<a href="Command-and-Function-Index.html#Command-and-Function-Index" title="Index" rel="index">Index</a>]</p>
</div>
<a name="Reduced-Form-Mapping-1"></a>
<h4 class="subsubsection">4.20.4.4 Reduced Form Mapping</h4>

<p>When the option <code>threshold_redform</code> is not set, or it is empty (the default), this analysis estimates a multivariate
smoothing spline ANOVA model (the &rsquo;mapping&rsquo;) for the selected entries in the transition matrix of the shock matrix of the reduce form first order solution of the model.
This mapping is done either with prior samples or with MC samples with <code>neighborhood_width</code>.
Unless <code>neighborhood_width</code> is set with MC samples, the  mapping of the reduced form solution forces the use of samples from
prior ranges or prior distributions, <i>i.e.</i>: <code>pprior=1</code> and <code>ppost=0</code>. 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.
</p>
<p>The following synthetic figures are produced:
</p><ul>
<li> <code>&lt;mod_file&gt;_redform_&lt;endo name&gt;_vs_lags_*.fig</code>: 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</code>) versus lagged endogenous variables (<code>namlagendo</code>); suffix
<code>log</code> indicates the results for log-transformed entries;

</li><li> <code>&lt;mod_file&gt;_redform_&lt;endo name&gt;_vs_shocks_*.fig</code>: 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</code>) versus exogenous variables (<code>namexo</code>); suffix <code>log</code>
indicates the results for log-transformed entries;

</li><li> <code>&lt;mod_file&gt;_redform_gsa(_log).fig</code>: 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.
</li></ul>

<p>Detailed results of the analyses are shown in the subfolder <code>&lt;mod_file&gt;/gsa/redform_prior</code> for prior samples and in <code>&lt;mod_file&gt;/gsa/redform_mc</code> for MC samples with option <code>neighborhood_width</code>,
where the detailed results of the estimation of the single functional relationships
between parameters <SPAN CLASS="MATH"><IMG
 WIDTH="12" HEIGHT="15" ALIGN="BOTTOM" BORDER="0"
 SRC="dynare.html_139.png"
 ALT="$\theta$"></SPAN> and reduced form coefficient (denoted as <SPAN CLASS="MATH"><IMG
 WIDTH="13" HEIGHT="30" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_42.png"
 ALT="$y$"></SPAN> hereafter) are stored in separate directories
named as:
</p>
<ul>
<li> <code>&lt;namendo&gt;_vs_&lt;namlagendo&gt;</code>: for the entries of the transition matrix;

</li><li> <code>&lt;namendo&gt;_vs_&lt;namexo&gt;</code>: for entries of the matrix of the shocks.
</li></ul>

<p>The following files are stored in each directory (we stick with prior sample but similar conventions are used for MC samples):
</p><ul>
<li> <code>&lt;mod_file&gt;_prior_&lt;namendo&gt;_vs_&lt;namexo&gt;.fig</code>: 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;

</li><li> <code>&lt;mod_file&gt;_prior_&lt;namendo&gt;_vs_&lt;namexo&gt;_map_SE.fig</code>: for entries of the shock matrix it shows graphs of the estimated first order ANOVA terms <!-- MATH
 $y = f(\theta_i)$
 -->
<SPAN CLASS="MATH"><IMG
 WIDTH="69" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_140.png"
 ALT="$y = f(\theta_i)$"></SPAN> for each deep parameter <SPAN CLASS="MATH"><IMG
 WIDTH="17" HEIGHT="30" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_141.png"
 ALT="$\theta_i$"></SPAN>; 

</li><li> <code>&lt;mod_file&gt;_prior_&lt;namendo&gt;_vs_&lt;namlagendo&gt;.fig</code>: 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;

</li><li> <code>&lt;mod_file&gt;_prior_&lt;namendo&gt;_vs_&lt;namlagendo&gt;_map_SE.fig</code>: for entries of the transition matrix it shows graphs of the estimated first order ANOVA terms <!-- MATH
 $y = f(\theta_i)$
 -->
<SPAN CLASS="MATH"><IMG
 WIDTH="69" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_140.png"
 ALT="$y = f(\theta_i)$"></SPAN> for each deep parameter <SPAN CLASS="MATH"><IMG
 WIDTH="17" HEIGHT="30" ALIGN="MIDDLE" BORDER="0"
 SRC="dynare.html_141.png"
 ALT="$\theta_i$"></SPAN>; 

</li><li> <code>&lt;mod_file&gt;_prior_&lt;namendo&gt;_vs_&lt;namexo&gt;_map.mat</code>, <code>&lt;mod_file&gt;_&lt;namendo&gt;_vs_&lt;namlagendo&gt;_map.mat</code>: these files store info in the estimation; 

</li></ul>

<p>When option <code>logtrans_redform</code> is set, the ANOVA estimation is performed using a log-transformation of each <code>y</code>.
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</code> suffix.
</p>
<p>When the option <code>threshold_redform</code> is set, the analysis is performed via Monte Carlo filtering, by displaying parameters that drive the individual entry <code>y</code> inside the range specified in <code>threshold_redform</code>. If no entry is found (or all entries are in the range), the MCF algorithm ignores the range specified in <code>threshold_redform</code> and performs the analysis splitting the MC sample of <code>y</code> into deciles. Setting <code>threshold_redform=[-inf inf]</code> triggers this approach for all <code>y</code>&rsquo;s.
</p>
<p>Results are stored in subdirectories of <code>&lt;mod_file&gt;/gsa/redform_prior</code> named 
</p><ul>
<li> <code>&lt;mod_file&gt;_prior_&lt;namendo&gt;_vs_&lt;namlagendo&gt;_threshold</code>: for the entries of the transition matrix;

</li><li> <code>&lt;mod_file&gt;_prior_&lt;namendo&gt;_vs_&lt;namexo&gt;_threshold</code>: for entries of the matrix of the shocks.
</li></ul>

<p>The files saved are named
</p><ul>
<li> <code>&lt;mod_file&gt;_prior_&lt;namendo&gt;_vs_&lt;namexo&gt;_threshold.fig</code>,<code>&lt;mod_file&gt;_&lt;namendo&gt;_vs_&lt;namlagendo&gt;_threshold.fig</code>: graphical outputs; 
</li><li> <code>&lt;mod_file&gt;_prior_&lt;namendo&gt;_vs_&lt;namexo&gt;_threshold.mat</code>,<code>&lt;mod_file&gt;_&lt;namendo&gt;_vs_&lt;namlagendo&gt;_threshold.mat</code>: info on the analysis; 

</li></ul>

<hr>
<a name="RMSE"></a>
<div class="header">
<p>
Next: <a href="#Screening-Analysis" accesskey="n" rel="next">Screening Analysis</a>, Previous: <a href="#Reduced-Form-Mapping" accesskey="p" rel="prev">Reduced Form Mapping</a>, Up: <a href="#Types-of-analysis-and-output-files" accesskey="u" rel="up">Types of analysis and output files</a> &nbsp; [<a href="index.html#SEC_Contents" title="Table of contents" rel="contents">Contents</a>][<a href="Command-and-Function-Index.html#Command-and-Function-Index" title="Index" rel="index">Index</a>]</p>
</div>
<a name="RMSE-1"></a>
<h4 class="subsubsection">4.20.4.5 RMSE</h4>

<p>The RMSE analysis can be performed with different types of sampling options:
</p><ol>
<li> When <code>pprior=1</code> and <code>ppost=0</code>, 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;

</li><li> When <code>pprior=0</code> and <code>ppost=0</code>, 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.</i> no Bayesian estimation);

</li><li> When <code>ppost=1</code> the toolbox analyzes the RMSEs for the posterior sample
obtained by Dynare&rsquo;s Metropolis procedure.
</li></ol>

<p>The use of cases 2 and 3 requires an estimation step beforehand. To
facilitate the sensitivity analysis after estimation, the <code>dynare_sensitivity</code>
command also allows you to indicate some options of the <code>estimation</code>
command. These are:
</p><ul>
<li> <code>datafile</code>
</li><li> <code>nobs</code>
</li><li> <code>first_obs</code>
</li><li> <code>prefilter</code>
</li><li> <code>presample</code>
</li><li> <code>nograph</code>
</li><li> <code>nodisplay</code>
</li><li> <code>graph_format</code>
</li><li> <code>conf_sig</code>
</li><li> <code>loglinear</code>
</li><li> <code>mode_file</code>
</li></ul>

<p>Binary files produced my RMSE analysis are:
</p><ul>
<li> <code>&lt;mod_file&gt;_prior_*.mat</code>: these files store the filtered and smoothed
    variables for the prior Monte-Carlo sample, generated when doing RMSE analysis
    (<code>pprior=1</code> and <code>ppost=0</code>);
</li><li> <code>&lt;mode_file&gt;_mc_*.mat</code>: these files store the filtered and smoothed variables
    for the multivariate normal Monte-Carlo sample, generated when doing
    RMSE analysis (<code>pprior=0</code> and <code>ppost=0</code>).
</li></ul>

<p>Figure files <code>&lt;mod_file&gt;_rmse_*.fig</code> store results for the RMSE analysis.
</p>
<ul>
<li> <code>&lt;mod_file&gt;_rmse_prior*.fig</code>: save results for the analysis using prior
Monte-Carlo samples;

</li><li> <code>&lt;mod_file&gt;_rmse_mc*.fig</code>: save results for the analysis using multivariate
normal Monte-Carlo samples;

</li><li> <code>&lt;mod_file&gt;_rmse_post*.fig</code>: save results for the analysis using Metropolis
posterior samples.
</li></ul>

<p>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):
</p>
<ul>
<li> <code>&lt;mod_file&gt;_rmse_prior_params_*.fig</code>: 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</code>);

</li><li> <code>&lt;mod_file&gt;_rmse_prior_&lt;var_obs&gt;_*.fig</code>: if a parameter significantly affects the fit of <code>var_obs</code>, all possible trade-off&rsquo;s with other observables for same parameter are plotted;

</li><li> <code>&lt;mod_file&gt;_rmse_prior_&lt;var_obs&gt;_map.fig</code>: plots the MCF analysis of parameters significantly driving the fit the observed series <code>var_obs</code>;

</li><li> <code>&lt;mod_file&gt;_rmse_prior_lnlik*.fig</code>: 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;

</li><li> <code>&lt;mod_file&gt;_rmse_prior_lnpost*.fig</code>: 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;

</li><li> <code>&lt;mod_file&gt;_rmse_prior_lnprior*.fig</code>: 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;

</li><li> <code>&lt;mod_file&gt;_rmse_prior_lik.fig</code>: when <code>lik_only=1</code>, this shows
the MCF tests for the filtering of the best 10% log-likelihood values;

</li><li> <code>&lt;mod_file&gt;_rmse_prior_post.fig</code>: when <code>lik_only=1</code>, this shows
the MCF tests for the filtering of the best 10% log-posterior values.
</li></ul>

<hr>
<a name="Screening-Analysis"></a>
<div class="header">
<p>
Next: <a href="#Identification-Analysis" accesskey="n" rel="next">Identification Analysis</a>, Previous: <a href="#RMSE" accesskey="p" rel="prev">RMSE</a>, Up: <a href="#Types-of-analysis-and-output-files" accesskey="u" rel="up">Types of analysis and output files</a> &nbsp; [<a href="index.html#SEC_Contents" title="Table of contents" rel="contents">Contents</a>][<a href="Command-and-Function-Index.html#Command-and-Function-Index" title="Index" rel="index">Index</a>]</p>
</div>
<a name="Screening-Analysis-1"></a>
<h4 class="subsubsection">4.20.4.6 Screening Analysis</h4>

<p>Screening analysis does not require any additional options with respect to
those listed in <a href="#Sampling-Options">Sampling Options</a>. The toolbox performs all the
analyses required and displays results.
</p>
<p>The results of the screening analysis with Morris sampling design are stored
in the subfolder <code>&lt;mod_file&gt;/gsa/screen</code>. The data file <code>&lt;mod_file&gt;_prior</code> stores
all the information of the analysis (Morris sample, reduced form coefficients,
etc.).
</p>
<p>Screening analysis merely concerns reduced form coefficients. Similar
synthetic bar charts as for the reduced form analysis with Monte-Carlo samples are
saved:
</p><ul>
<li> <code>&lt;mod_file&gt;_redform_&lt;endo name&gt;_vs_lags_*.fig</code>: 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</code>) versus lagged endogenous variables (<code>namlagendo</code>);

</li><li> <code>&lt;mod_file&gt;_redform_&lt;endo name&gt;_vs_shocks_*.fig</code>: 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</code>) versus exogenous variables (<code>namexo</code>);

</li><li> <code>&lt;mod_file&gt;_redform_screen.fig</code>: 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.
</li></ul>

<hr>
<a name="Identification-Analysis"></a>
<div class="header">
<p>
Previous: <a href="#Screening-Analysis" accesskey="p" rel="prev">Screening Analysis</a>, Up: <a href="#Types-of-analysis-and-output-files" accesskey="u" rel="up">Types of analysis and output files</a> &nbsp; [<a href="index.html#SEC_Contents" title="Table of contents" rel="contents">Contents</a>][<a href="Command-and-Function-Index.html#Command-and-Function-Index" title="Index" rel="index">Index</a>]</p>
</div>
<a name="Identification-Analysis-1"></a>
<h4 class="subsubsection">4.20.4.7 Identification Analysis</h4>

<p>Setting the option <code>identification=1</code>, 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.
</p>
<p>Prerequisite for properly running all the identification routines, is the keyword
<code>identification</code>; 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</code>,
which implements <cite>Iskrev (2010)</cite> identification analysis.
</p>
<p>For example, the placing <code>identification; dynare_sensitivity(identification=1, morris=2);</code>
in the Dynare model file trigger identification analysis using analytic derivatives
<cite>Iskrev (2010)</cite>, jointly with the mapping of the acceptable region.
</p>
<p>The identification analysis with derivatives can also be triggered by the
commands <code>identification;</code> 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.
</p>

<hr>
<div class="header">
<p>
Previous: <a href="#Screening-Analysis" accesskey="p" rel="prev">Screening Analysis</a>, Up: <a href="#Types-of-analysis-and-output-files" accesskey="u" rel="up">Types of analysis and output files</a> &nbsp; [<a href="index.html#SEC_Contents" title="Table of contents" rel="contents">Contents</a>][<a href="Command-and-Function-Index.html#Command-and-Function-Index" title="Index" rel="index">Index</a>]</p>
</div>



</body>
</html>
dynare-doc 4.5.4-1 / usr / share / doc / dynare / dynare.html / Sensitivity-and-identification-analysis.html
