stepan
/
preprocessor
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

macroprocessor.tex: update slides

fix-tolerance-parameters
Johannes Pfeifer 2022-04-11 23:42:37 +02:00
parent 71edfd05e4
commit e85d085ae8
1 changed files with 83 additions and 52 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

135
doc/macroprocessor/macroprocessor.tex
View File

@ -1,7 +1,8 @@
\documentclass{beamer}
\documentclass[aspectratio=169]{beamer}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage{lmodern}
\usepackage{upquote}
\usepackage{amsmath}
\usepackage[copyright]{ccicons}
@ -15,7 +16,7 @@
\institute[CEPREMAP]{\includegraphics[scale=0.15]{../logos/cepremap.jpg}}
\date{21 June 2021}
\date{11 April 2022}
\setbeamertemplate{title page}
@ -66,7 +67,7 @@
\ccbysa
\column{0.71\textwidth}
\tiny
Copyright © 2008--2021 Dynare Team \\
Copyright © 2008--2022 Dynare Team \\
Licence: \href{http://creativecommons.org/licenses/by-sa/4.0/}{Creative
Commons Attribution-ShareAlike 4.0}
\end{columns}
@ -90,7 +91,7 @@
\item The \textbf{Dynare macro language} adds a programmatic element to Dynare
\begin{itemize}
\item Introduces conditionals, loops, and other simple programmatic directives
\item Used to speed model development
\item Used to speed up model development
\item Useful in various situations
\begin{itemize}
\item Multi-country models
@ -132,6 +133,7 @@
\item loop statements: \verb+@#for/@#endfor+
\end{itemize}
\item Most directives fit on one line. If needed however, two backslashes (i.e. \verb+\\+) at the end of a line indicate that the directive is continued on the next line.
\item Directives are not terminated with a semicolon
\end{itemize}
\end{frame}
@ -139,13 +141,13 @@
\frametitle{Values}
\begin{itemize}
\item The macro processor can handle values of 5 different types:
\begin{itemize}
\begin{enumerate}
\item boolean
\item real
\item string
\item tuple
\item array
\end{itemize}
\end{enumerate}
\item Values of the types listed above can be cast to other types
\begin{itemize}
\item \texttt{(real) "3.1"} $\rightarrow$ \texttt{3.1}
@ -161,12 +163,12 @@
\begin{frame}[fragile=singleslide]
\frametitle{Macro-expressions (1/8)}
\begin{itemize}
\item Macro-expressions are constructed using literals of the 5 basic types
\item Macro-expressions are constructed using literals (i.e.\ fixed values) of the 5 basic types
described above, macro-variables, standard operators, function calls and comprehensions.
\item Macro-expressions can be used in two places:
\begin{itemize}
\item inside macro directives; no special markup is required
\item in the body of the \texttt{.mod} file, between an at sign and curly braces (like \verb+@{expr}+); the macro processor will substitute the expression with its value
\item in the body of the \texttt{.mod} file, between an ``at''-sign and curly braces (like \verb+@{expr}+); the macro processor will substitute the expression with its value
\end{itemize}
\end{itemize}
\end{frame}
@ -223,7 +225,7 @@
\begin{frame}[fragile=singleslide]
\frametitle{Macro-expressions (5/8): Tuple}
Tuples are enclosed by parenthesis and elements separated by commas (like
Tuples are enclosed by parentheses and elements are separated by commas (like
\texttt{(a,b,c)} or \texttt{(1,2.2,c)}).
\begin{block}{Operators on tuples}
\begin{itemize}
@ -256,7 +258,7 @@
\frametitle{Macro-expressions (6/8): Array (2/2)}
Arrays can be seen as representing a set of elements (assuming no element
appears twice in the array). Several set operations can thus be performed on
arrays: union, intersection, difference, cartesian product and power.
arrays: union, intersection, difference, Cartesian product and power.
\begin{block}{Set operations on arrays}
\begin{itemize}
\item set union: \texttt{|}
@ -266,7 +268,7 @@
\item Cartesian power of an array: \texttt{\^}
\end{itemize}
\end{block}
For example, if \texttt{A} and \texttt{B} are arrays, then the following
For example: if \texttt{A} and \texttt{B} are arrays, then the following
set operations are valid: \texttt{A|B}, \texttt{A\&B}, \texttt{A-B},
\texttt{A*B}, \texttt{A\^{}3}.
\end{frame}
@ -329,7 +331,7 @@
\frametitle{Macro-expressions (8/8): Functions}
\begin{itemize}
\item Can take any number of arguments
\item Dynamic binding: is evaluated when invoked, not when defined
\item Dynamic binding: is evaluated when invoked during the macroprocessing stage, not when defined
\item Can be included in expressions; valid operators depend on return type
\end{itemize}
@ -352,7 +354,7 @@ Then \texttt{distance(3, 4)} will be equivalent to \texttt{5}.
The value of a macro-variable can be defined with the \verb+@#define+
directive.
The macro processor has its own list of variables which are different from model variables and MATLAB/Octave variables
The macro processor has its own list of variables, which are different from model variables and MATLAB/Octave variables
\begin{block}{Syntax}
\verb+@#define +\textit{variable\_name}\verb+ = +\textit{expression}
@ -368,7 +370,7 @@ Then \texttt{distance(3, 4)} will be equivalent to \texttt{5}.
@#define t = ("US" in w) // Equals true
\end{verbatim}
\end{block}
NB: You can define macro variables on the dynare command line by using the \texttt{-D} option
NB: You can define macro variables on the Dynare command line by using the \texttt{-D} option
\end{frame}
\begin{frame}[fragile=singleslide]
@ -427,7 +429,7 @@ end;
\item Files to include are searched for in the current directory. Other directories can
be added with the
\verb+@#includepath+ directive, the \texttt{-I} command line option, or the
\texttt{[paths]} section in config file.
\texttt{[paths]} section in config files.
\end{itemize}
\end{frame}
@ -539,7 +541,8 @@ This loop will iterate over only 4 triplets: \texttt{(3,4,5)},
\begin{frame}[fragile=singleslide]
\frametitle{Conditional directives (2/3)}
\begin{block}{Syntax 3}
\verb+@#if +\textit{bool\_or\_real\_expr1} \\
\scriptsize
\verb+@#if +\textit{bool\_or\_real\_expr1} \\
\verb+ +\textit{body included if expr1 is true (or != 0)} \\
\verb+@#elseif +\textit{bool\_or\_real\_expr2} \\
\verb+ +\textit{body included if expr2 is true (or != 0)} \\
@ -588,35 +591,53 @@ end;
\end{columns}
\bigskip
There is also \verb+@#ifndef+, which is the opposite of \verb+@#ifdef+
(\textit{i.e.} it tests whether a variable is \emph{not} defined). NB: There is
\begin{itemize}
\item There is also \verb+@#ifndef+, which is the opposite of \verb+@#ifdef+
(\textit{i.e.} it tests whether a variable is \emph{not} defined).
\item NB: There is
\emph{no} \verb+@#elseifdef+ or \verb+@#elseifndef+ directive; use
\verb+elseif defined(variable_name)+ to achieve the desired objective.
\end{itemize}
\end{frame}
\begin{frame}[fragile=singleslide]
\frametitle{Echo and error directives}
\frametitle{Echo directives}
\begin{itemize}
\item The echo directive will simply display a message on standard output
\item The error directive will display the message and make Dynare stop (only makes sense inside a conditional directive)
\item The echomacrovars directive will display all of the macro variables (or
those specified) and their values, optionally saving them
those specified) and their values
\item The \texttt{save} option allows saving this information to \texttt{options\_.macrovars\_line\_x}, where \texttt{x} denotes the line number where the statement was encountered
\end{itemize}
\begin{block}{Syntax}
\verb+@#echo +\textit{string\_expr} \\
\verb+@#error +\textit{string\_expr} \\
\verb+@#echomacrovars +\\
\verb+@#echomacrovars +\textit{list\_of\_variables}\\
\verb+@#echomacrovars(save)+\\
\verb+@#echomacrovars(save) +\textit{list\_of\_variables}\\
\verb+@#echo +\textit{string\_expr} \\
\verb+@#echomacrovars +\\
\verb+@#echomacrovars +\textit{list\_of\_variables}\\
\verb+@#echomacrovars(save)+\\
\verb+@#echomacrovars(save) +\textit{list\_of\_variables}\\
\end{block}
\begin{block}{Examples}
\begin{verbatim}
@#echo "Information message."
\end{verbatim}
\end{block}
\end{frame}
\begin{frame}[fragile=singleslide]
\frametitle{Error directive}
\begin{itemize}
\item The error directive will display the message and make Dynare stop (only makes sense inside a conditional directive)
\end{itemize}
\begin{block}{Syntax}
\verb+@#error +\textit{string\_expr} \\
\end{block}
\begin{block}{Example}
\begin{verbatim}
@#error "Error message!"
\end{verbatim}
\end{block}
@ -626,8 +647,8 @@ There is also \verb+@#ifndef+, which is the opposite of \verb+@#ifdef+
\frametitle{Macro-related command line options}
\begin{itemize}
\item \texttt{savemacro}: Useful for debugging or learning purposes, saves the output of the macro processor. If your \texttt{.mod} file is called \texttt{file.mod}, the output is saved to \texttt{file-macroexp.mod}.
\item \texttt{nolinemacro}: In the output of \texttt{savemacro}, don't print line numbers where the macro directives were placed.
\item \texttt{noemptylinemacro}: Remove empty lines in the output of \texttt{savemacro}.
\item NB: \texttt{savemacro=filename} allows a user-defined file name
\item \texttt{linemacro}: In the output of \texttt{savemacro}, print line numbers where the macro directives were placed.
\item \texttt{onlymacro}: Stops processing after the macro processing step.
\end{itemize}
\end{frame}
@ -640,13 +661,13 @@ There is also \verb+@#ifndef+, which is the opposite of \verb+@#ifdef+
\item The \verb+@#include+ directive can be used to split \texttt{.mod} files into several modular components
\item Example setup:
\begin{description}
\item[\texttt{modeldesc.mod}:] contains variable declarations, model equations and shocks declarations
\item[\texttt{simulate.mod}:] includes \texttt{modeldesc.mod}, calibrates parameters and runs stochastic simulations
\item[\texttt{estim.mod}:] includes \texttt{modeldesc.mod}, declares priors on parameters and runs bayesian estimation
\item[\texttt{modeldesc.mod}:] contains variable declarations, model equations, and shock declarations
\item[\texttt{simulate.mod}:] includes \texttt{modeldesc.mod}, calibrates parameters, and runs stochastic simulations
\item[\texttt{estim.mod}:] includes \texttt{modeldesc.mod}, declares priors on parameters, and runs Bayesian estimation
\end{description}
\item Dynare can be called on \texttt{simulate.mod} and \texttt{estim.mod}
\item But it makes no sense to run it on \texttt{modeldesc.mod}
\item Advantage: no need to manually copy/paste the whole model (at the beginning) or changes to the model (during development)
\item Advantage: no need to manually copy/paste the whole model (during initial development) or port model changes (during development)
\end{itemize}
\end{frame}
@ -725,18 +746,18 @@ end;
\begin{frame}
\frametitle{Endogeneizing parameters (1/4)}
\begin{itemize}
\item When calibrating the model it may be useful to consider a parameter as an endogenous (and vice-versa)
\item When calibrating the model, it may be useful to pin down parameters by targeting endogenous objects
\item Example:
\begin{gather*}
y = \left(\alpha^{\frac{1}{\xi}} \ell^{1-\frac{1}{\xi}} + (1-\alpha)^{\frac{1}{\xi}}k^{1-\frac{1}{\xi}}\right)^{\frac{\xi}{\xi - 1}} \\
lab\_rat = \frac{w \ell}{p y}
y_t = \left(\alpha^{\frac{1}{\xi}} \ell_t^{1-\frac{1}{\xi}} + (1-\alpha)^{\frac{1}{\xi}}k_t^{1-\frac{1}{\xi}}\right)^{\frac{\xi}{\xi - 1}} \\
lab\_rat_t = \frac{w_t \ell_t}{p_t y_t}
\end{gather*}
\item In the model, $\alpha$ is a (share) parameter, and $lab\_rat$ is an endogenous variable
\item In the model, $\alpha$ is a (share) parameter, and $lab\_rat_t$ is an endogenous variable
\item We observe that:
\begin{itemize}
\item calibrating $\alpha$ is not straigthforward!
\item on the contrary, we have real world data for $lab\_rat$
\item it is clear that these two variables are economically linked
\item setting a value for $\alpha$ is not straightforward!
\item but we have real world data for $lab\_rat_t$
\item it is clear that these two objects are economically linked
\end{itemize}
\end{itemize}
\end{frame}
@ -744,13 +765,13 @@ end;
\begin{frame}[fragile=singleslide]
\frametitle{Endogeneizing parameters (2/4)}
\begin{itemize}
\item Therefore, when computing the steady state:
\item Therefore, when computing the steady state by solving the static model:
\begin{itemize}
\item we make $\alpha$ an endogenous variable and $lab\_rat$ a parameter
\item we impose an economically relevant value for $lab\_rat$
\item we make $\alpha$ a variable and the steady state value $lab\_rat$ of the dynamic variable $lab\_rat_t$ a parameter
\item we impose an economically sensible value for $lab\_rat$
\item the solution algorithm deduces the implied value for $\alpha$
\end{itemize}
\item We call this method ``variable flipping''
\item We call this method ``variable flipping'', because it treats $\alpha$ as a variable and $lab\_rat$ as a parameter for the purpose of the static model
\end{itemize}
\end{frame}
@ -785,16 +806,16 @@ end;
\item File \texttt{steadystate.mod}:
\begin{itemize}
\item begins with \verb+@#define steady = true+
\item then with \verb+@#include "modeqs.mod"+
\item followed by \verb+@#include "modeqs.mod"+
\item initializes parameters (including \texttt{lab\_rat}, excluding \texttt{alpha})
\item computes steady state (using guess values for endogenous, including \texttt{alpha})
\item saves values of parameters and endogenous at steady-state in a file, using the \texttt{save\_params\_and\_steady\_state} command
\item saves values of parameters and variables at steady-state in a file, using the \texttt{save\_params\_and\_steady\_state} command
\end{itemize}
\item File \texttt{simulate.mod}:
\begin{itemize}
\item begins with \verb+@#define steady = false+
\item then with \verb+@#include "modeqs.mod"+
\item loads values of parameters and endogenous at steady-state from file, using the \texttt{load\_params\_and\_steady\_state} command
\item followed by \verb+@#include "modeqs.mod"+
\item loads values of parameters and variables at steady-state from file, using the \texttt{load\_params\_and\_steady\_state} command
\item computes simulations
\end{itemize}
\end{itemize}
@ -809,14 +830,18 @@ end;
\begin{verbatim}
rhos = [ 0.8, 0.9, 1];
for i = 1:length(rhos)
rho = rhos(i);
set_param_value('rho',rhos(i));
stoch_simul(order=1);
if info(1)~=0
error('Simulation failed for parameter draw')
end
end
\end{verbatim}
\end{block}
\begin{itemize}
\item The loop is not unrolled
\item MATLAB/Octave manages the iterations
\item NB: always check whether the error flag `info(1)==0` to prevent erroneously relying on stale results from previous iterations
\end{itemize}
\end{frame}
@ -826,8 +851,11 @@ end
\begin{verbatim}
rhos = [ 0.8, 0.9, 1];
@#for i in 1:3
rho = rhos(@{i});
set_param_value('rho',rhos(@{i}));
stoch_simul(order=1);
if info(1)~=0
error('Simulation failed for parameter draw')
end
@#endfor
\end{verbatim}
\end{block}
@ -843,8 +871,11 @@ rhos = [ 0.8, 0.9, 1];
\begin{block}{With a macro processor loop (case 2)}
\begin{verbatim}
@#for rho_val in [ 0.8, 0.9, 1]
rho = @{rho_val};
set_param_value('rho',@{rho_val});
stoch_simul(order=1);
if info(1)~=0
error('Simulation failed for parameter draw')
end
@#endfor
\end{verbatim}
\end{block}