Math & Science ⇒ Problems with 'tcilatex' Macros

[lucianojcampos]

Hi, I can´t run my tex file when I use

\input{tcilatex}

even though I´ve already downloaded the tcilatex.tex latest version (2002-09-30) from:

http://www.filewatcher.com/m/tcilatex.t ... 2.0.0.html

I appreciate any help you could gave me.

Thanks,

[sitex]

What happens if you remove the \input{tcilatex}? Normally, you do not need this command unless you have something like %TCIMACRO{... your file. How did you create your latex file? Please post a minimal (nonworking) example.

Regards,

Tom

[lucianojcampos]

Hi Tom, thanks for your answer. Even when I remove \input{tcilatex} I still have problems to run it. More specifically, I get a warning sign telling me that the following package is not found:

\usepackage{sw20elba}

I post here the complete tex file

I really appreciate your help.
Best,

Code: Select all

%2multibyte Version: 5.50.0.2953 CodePage: 1252

\documentclass[12pt]{article}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\usepackage{amsfonts}
\usepackage{amsmath}
\usepackage{sw20elba}
\usepackage{harvard}
\usepackage{amsbsy}
\usepackage{portland}
\usepackage[pdftex]{hyperref}

\setcounter{MaxMatrixCols}{10}
%TCIDATA{OutputFilter=LATEX.DLL}
%TCIDATA{Version=5.50.0.2953}
%TCIDATA{Codepage=1252}
%TCIDATA{<META NAME="SaveForMode" CONTENT="1">}
%TCIDATA{BibliographyScheme=BibTeX}
%TCIDATA{Created=Tuesday, December 03, 2002 15:46:39}
%TCIDATA{LastRevised=Sunday, April 03, 2011 15:52:31}
%TCIDATA{<META NAME="GraphicsSave" CONTENT="32">}
%TCIDATA{<META NAME="DocumentShell" CONTENT="Articles\SW\Elbert Walker's">}
%TCIDATA{CSTFile=LaTeX article (bright).cst}

\newtheorem{theorem}{Theorem}
\newtheorem{acknowledgement}[theorem]{Acknowledgement}
\newtheorem{algorithm}[theorem]{Algorithm}
\newtheorem{axiom}[theorem]{Axiom}
\newtheorem{case}[theorem]{Case}
\newtheorem{claim}[theorem]{Claim}
\newtheorem{conclusion}[theorem]{Conclusion}
\newtheorem{condition}[theorem]{Condition}
\newtheorem{conjecture}[theorem]{Conjecture}
\newtheorem{corollary}[theorem]{Corollary}
\newtheorem{criterion}[theorem]{Criterion}
\newtheorem{definition}[theorem]{Definition}
\newtheorem{example}[theorem]{Example}
\newtheorem{exercise}[theorem]{Exercise}
\newtheorem{lemma}[theorem]{Lemma}
\newtheorem{notation}[theorem]{Notation}
\newtheorem{problem}[theorem]{Problem}
\newtheorem{proposition}[theorem]{Proposition}
\newtheorem{remark}[theorem]{Remark}
\newtheorem{solution}[theorem]{Solution}
\newtheorem{summary}[theorem]{Summary}
\newenvironment{proof}[1][Proof]{\textbf{#1.} }{\ \rule{0.5em}{0.5em}}
\renewcommand{\cite}{\citeasnoun}
\input{tcilatex}
\begin{document}

\title{Decomposing changes in the transmission mechanism of structural
shocks since the 1970}
\author{Luciano Campos, PhD in Economics programme, M\'{e}moire de recherche}
\date{}
\maketitle

\begin{abstract}
{\small Xxxxxxxxxxx.}
\end{abstract}

\section{Introduction}

[Two questions explored\ with this memoire:

\qquad (1)\ has the transmission mechanism of structural shocks (both policy
and non-policy) changed since the Great Inflation period?\ 

\qquad (2)\ If the answer to (1)\ is Yes, what accounts for such changes?
Changes in the structure of the private sector, in the systematic component
of monetary policy, in the disturbances impacting upon the economy, or in
the level of trend inflation?]

[Within a DSGE\ context, such investigation has, so far, neglected the
possible role of trend inflation, which, in the U.S., was around 7-8 per
cent during those years (see the estimates from the work of Cogley and
Sargent), so non negligible. (Here some references on previous work on this
based on DSGE\ models).]

[The work of Guido Ascari, however, has shown that New Keynesian models
log-linearised around non-zero trend inflation have a more complex structure
than the one you obtain when you log-linearised such models around zero
trend inflation. This implies that the estimates which exist in the
literature---which have been systematically obtained based on models
log-linearised around zero trend inflaiton---may well be distorted, and may
therefore provide a distorted picture.]

[In this memoire I\ estimate the New Keynesian model of \cite{AscariRopele},
which generalises the standard New Keynesian model analysed by\ \cite%
{cgg2000}\ and \cite{woodfordinterestprices}\ to the case of non-zero trend
inflation, for the Great Inflation period and for the period between the end
of the Volcker disinflation and the beginning of the financial crisis, and I
analyse how several features of the model---the private sector parameters,
the systematic component of monetary policy, the structural disturbances,
and trend inflation---account for the changes in the response of the economy
to structural shocks between the two periods.]

[Here discuss the results]

The memoire is structured as follows. [Here finish]

\section{The Model}

The\ model\ we use is\ the\ one proposed by\ \cite{AscariRopele}, which
generalises the standard New Keynesian model analysed by\ \cite{cgg2000}\
and \cite{woodfordinterestprices}\ to the case of non-zero trend inflation,
nesting it as a particular case.

The Phillips curve block of the model is given by%
\begin{equation}
\Delta _{t}=\psi \Delta _{t+1|t}+\eta \phi _{t+1|t}+\kappa \frac{\sigma _{N}%
}{1\text{+}\sigma _{N}}s_{t}+\kappa y_{t}+\epsilon _{\pi ,t}  \label{Ascari1}
\end{equation}%
\begin{equation}
\phi _{t}=\chi \phi _{t+1|t}+\chi (\theta \text{-1})\Delta _{t+1|t}
\label{Ascari2}
\end{equation}%
\begin{equation}
s_{t}=\xi \Delta _{t}+\alpha \bar{\pi}^{\theta (\text{1-}\epsilon )}s_{t-1}
\label{Ascari3}
\end{equation}%
where $\Delta _{t}\equiv \pi _{t}$-$\tau \epsilon \pi _{t-1}$;\ $\pi _{t}$, $%
y_{t}$, and $s_{t}$\ are\ the log-deviations of inflation, the output gap,
and the dispersion of relative prices,\ respectively, from the
non-stochastic\ steady-state;\ $\theta $\TEXTsymbol{>}1 is the elasticity
parameter in the aggregator function turning intermediate inputs into the
final good; $\alpha $ is the Calvo parameter;\ $\epsilon \in $[0,1]\ is the
degree of indexation;\ $\tau \in $[0,1]\ parameterises\ the extent to which
indexation is to past inflation as opposed to trend inflation\ (with $\tau $%
=1 indexation is to past inflation, whereas with $\tau $=0 indexation is to
trend inflation);\ $\Delta _{t}$\ and $\phi _{t}$\ are auxiliary variables; $%
\sigma _{N}$ is the inverse of the elasticity of intertemporal substitution
of labor, which, following \cite{AscariRopele},\ we calibrate to 1; and\ $%
\psi \equiv \beta \bar{\pi}^{1-\epsilon }$+$\eta (\theta $-1$)$, $\chi
\equiv \alpha \beta \bar{\pi}^{(\theta \text{-1})(\text{1-}\epsilon )}$, $%
\xi \equiv (\bar{\pi}^{1-\epsilon }$-1$)\theta \alpha \bar{\pi}^{(\theta 
\text{-1})(\text{1-}\epsilon )}[$1-$\alpha \bar{\pi}^{(\theta \text{-1})(%
\text{1-}\epsilon )}]^{-1}$, $\eta \equiv \beta (\bar{\pi}^{1-\epsilon }$-1$%
)[$1-$\alpha \bar{\pi}^{(\theta \text{-1})(\text{1-}\epsilon )}]$, and $%
\kappa \equiv $($1$+$\sigma _{N}$)$[\alpha \bar{\pi}^{(\theta \text{-1})(%
\text{1-}\epsilon )}]^{-1}[$1-$\alpha \beta \bar{\pi}^{\theta (\text{1-}%
\epsilon )}][$1-$\alpha \bar{\pi}^{(\theta \text{-1})(\text{1-}\epsilon )}]$%
, where $\bar{\pi}$\ is gross trend inflation measured on a
quarter-on-quarter basis.\footnote{%
To be clear, this implies that\ (e.g.) a steady-state inflation rate of 4
per cent per year maps into a value of $\bar{\pi}$\ equal to 1.04$^{\text{1/4%
}}$=1.00985.} In what follows we uniquely consider the case of indexation to 
\textit{past} inflation, and we therefore set $\tau $=1. We close the\ model
with the intertemporal IS\ curve%
\begin{equation}
y_{t}=\gamma y_{t+1|t}+(1-\gamma )y_{t-1}-\sigma ^{-1}(R_{t}-\pi
_{t+1|t})+\epsilon _{y,t}  \label{Ascari4}
\end{equation}%
and the monetary policy rule%
\begin{equation}
R_{t}=\rho R_{t-1}+(1-\rho )[\phi _{\pi }\pi _{t}+\phi _{y}y_{t}]+\epsilon
_{R,t}  \label{Ascari5}
\end{equation}

\section{Bayesian Estimation}

We\ estimate (\ref{Ascari1})-(\ref{Ascari5})\textit{\ via} Bayesian
methods.\ The next two-sub-appendices describe the\ priors, and the
Markov-Chain Monte Carlo algorithm we use to get draws from the posterior.

\subsection{Priors}

Following,\ e.g.,\ \cite{lubschorf2002}\ and \cite{anschorfheide},\ all
structural parameters are assumed, for the sake of simplicity, to be\ 
\textit{a priori}\ independent from one another.\ Table 1\ reports the
parameters'\ prior densities,\ together with two key objects\ characterising
them, the mode and the standard deviation.

\subsection{Numerical\ maximisation of\ the log posterior}

We\ numerically\ maximise\ the log posterior---defined as ln $L$($\theta $%
\TEXTsymbol{\vert}$Y$) + ln $P$($\theta $), where $\theta $ is the vector
collecting\ the model's structural parameters,\ $L$($\theta $\TEXTsymbol{%
\vert}$Y$) is the likelihood of $\theta $ conditional on the data, and $P$($%
\theta $)\ is the prior---\textit{via}\ simulated\ annealing.\ Following 
\cite{goffeferrierrogers} we implement simulated\ annealing \textit{via} the
algorithm proposed by\ \cite{coranasimulatedannealing}, setting the key
parameters to $T_{0}$=100,000, $r_{T}$=0.9, $N_{t}$=5, $N_{s}$=20, $\epsilon 
$=10$^{-6}$, $N_{\epsilon }$=4, where\ $T_{0}$\ is the initial temperature, $%
r_{T}$ is the\ temperature reduction factor, $N_{t}$\ is the\ number of
times the algorithm goes through the\ $N_{s}$ loops\ before the temperature
starts being reduced,\ $N_{s}$\ is the\ number of times the algorithm goes
through the\ function before adjusting the stepsize, $\epsilon $\ is the
convergence (tolerance) criterion,\ and $N_{\epsilon }$\ is number of times
convergence\ is achieved before the algorithm stops. Finally, initial
conditions were chosen stochastically by the algorithm itself, while the
maximum number of functions evaluations,\ set\ to 1,000,000,\ was not
achieved.

\subsection{Getting\ draws\ from\ the posterior\ \textit{via} Random-Walk
Metropolis}

We generate\ draws from the posterior distribution of the model's structural
parameters\ \textit{via} the Random Walk Metropolis (henceforth, RWM)
algorithm as described in, e.g.,\ \cite{anschorfheide}. In implementing the
RWM algorithm we exactly follow An and Schorfheide (2007, Section 4.1), with
the single exception of the method\ we use to calibrate the covariance
matrix's scale factor---the parameter \textit{c} below---for which we follow
the\ methodology described\ in Appendix D.2 of \cite%
{BenatiInflationPersistenceQJE} in order to get a fraction of accepted draws
close to the ideal one (in high dimensions)\ of 0.23.\footnote{%
See \cite{gelmancarlinsternrubin}.}

Let then\ $\hat{\theta}$\ and $\hat{\Sigma}$\ be the mode of the maximised\
log\ posterior and its\ estimated Hessian, respectively.\footnote{%
We compute $\hat{\Sigma}$ numerically as in \cite{anschorfheide}.} We start
the Markov chain of the\ RWM algorithm by drawing\ $\theta ^{(0)}$\ from $N$(%
$\hat{\theta}$, $c^{2}\hat{\Sigma}$).\ For $s$ = 1, 2, ..., $N$ we then draw 
$\tilde{\theta}$\ from the proposal distribution $N$($\theta ^{(s-1)}$, $%
c^{2}\hat{\Sigma}$), accepting the jump (i.e., $\theta ^{(s)}$\ = $\tilde{%
\theta}$) with probability min \{1, $r$($\theta ^{(s-1)}$, $\theta $%
\TEXTsymbol{\vert}$Y$)\}, and rejecting it (i.e., $\theta ^{(s)}$\ = $\theta
^{(s-1)}$) otherwise, where%
\begin{equation*}
r(\theta ^{(s-1)},\theta |Y)=\frac{L(\theta |Y)\text{ }P(\theta )}{L(\theta
^{(s-1)}|Y)\text{ }P(\theta ^{(s-1)})}
\end{equation*}%
We\ run a burn-in sample of\ 200,000 draws which\ we then discard. After
that,\ we run a sample of 500,000 draws, keeping every draw out of 100\ in
order to decrease the draws' autocorrelation.

\subsection{Handling the possibility of indeterminacy in estimation}

An important issue in estimation concerns how to handle the possibility of
indeterminacy. In a string of papers,\footnote{%
See in particular \cite{AscariRED2004}\ and \cite{AscariRopele}.}\ Guido
Ascari has indeed shown that, when standard New Keynesian models are\
log-linearised\ around a non-zero steady-state inflation\ rate, the size of
the determinacy region is, for\ a given parameterisation,\ `shrinking'\
(i.e., decreasing)\ in the level of trend inflation.\footnote{%
On this, see also\ \cite{KileyTrendInflation}.}\ \cite{AscariRopele} in
particular\ show that, conditional on their calibration, it is very\
difficult to obtain a determinate equilibrium for values of trend inflation
beyond 4 to 6 per cent. Given that, for all of the countries in our sample,
inflation has been beyond this threshold\ for a significant portion of the
sample period\ (first and foremost, during\ the Great Inflation episode),\
the imposition of determinacy in estimation over the entire sample\ would be
very\ hard to justify.\ In what follows we therefore estimate the model
given by (\ref{Ascari1})-(\ref{Ascari5}) by allowing for the possibility of 
\textit{one-dimensional} indeterminacy,\footnote{%
This is in line with \cite{justinianoprimiceri}. As they stress (see Section
8.2.1), `[t]his means that we effectively truncate our prior at the boundary
of a multi-dimensional indeterminacy region'.} and further imposing the
constraint that, when trend inflation is lower than 3 per cent, the economy
is within the determinacy region.\footnote{%
The constraint that,\ below 3 per cent trend inflation, the economy is under
determinacy was\ imposed in order to rule out a few highly implausible
estimates we obtained when no such constraint was imposed. In particular,
without imposing any constraint, in a few cases estimates would point
towards the economy being under indeterminacy even within the current
low-inflation environment, which we find \textit{a priori} hard to believe.
These results originate from the fact that, as stressed e.g. by \cite%
{lubschorf2002}, (in)determinacy is a \textit{system} property, crucially
depending on the interaction between \textit{all} of the (policy or
non-policy) structural parameters, so that parameters'\ configurations\
which, within the comparatively simple New Keynesian model used herein,
produce the best fit to the data may produce such undesirable\ `side
effects'.}

\bigskip 

\newpage 

\section{Conclusions}

[...].

\newpage

\bibliographystyle{ECONOMETRICA}
\bibliography{ACOMPAT,Multi}

\newpage

\bigskip 

\bigskip 

\bigskip 

\begin{tabular}{||c|c|c|cc||}
\hline\hline
\multicolumn{5}{||l||}{\textbf{Table 1\ \ Prior distributions for the New}}
\\ 
\multicolumn{5}{||l||}{\textbf{Keynesian\ model's structural parameters}} \\ 
\hline
&  &  &  & {\small Standard} \\ 
{\small Parameter} & {\small Domain} & {\small Density} & {\small Mode} & 
{\small deviation} \\ \hline
$\theta $-1 & $%
%TCIMACRO{\U{211d} }%
%BeginExpansion
\mathbb{R}
%EndExpansion
^{+}$ & {\small Gamma} & {\small 10} & {\small 5} \\ 
$\alpha $ & {\small [0, 1)} & {\small Beta} & {\small 0.588} & {\small 0.02}
\\ 
$\epsilon $ & {\small [0, 1]} & {\small Uniform} & {\small --} & {\small %
0.2887} \\ 
$\sigma $ & $%
%TCIMACRO{\U{211d} }%
%BeginExpansion
\mathbb{R}
%EndExpansion
^{+}$ & {\small Gamma} & {\small 2} & {\small 1} \\ 
$\delta $ & {\small [0, 1]} & {\small Uniform} & {\small --} & {\small 0.2887%
} \\ 
$\sigma _{R}^{2}$ & $%
%TCIMACRO{\U{211d} }%
%BeginExpansion
\mathbb{R}
%EndExpansion
^{+}$ & {\small Inverse Gamma} & {\small 0.5} & {\small 5} \\ 
$\sigma _{\pi }^{2}$ & $%
%TCIMACRO{\U{211d} }%
%BeginExpansion
\mathbb{R}
%EndExpansion
^{+}$ & {\small Inverse Gamma} & {\small 0.5} & {\small 5} \\ 
$\sigma _{y}^{2}$ & $%
%TCIMACRO{\U{211d} }%
%BeginExpansion
\mathbb{R}
%EndExpansion
^{+}$ & {\small Inverse Gamma} & {\small 0.5} & {\small 5} \\ 
$\sigma _{s}^{2}$ & $%
%TCIMACRO{\U{211d} }%
%BeginExpansion
\mathbb{R}
%EndExpansion
^{+}$ & {\small Inverse Gamma} & {\small 0.1} & {\small 0.1} \\ 
$\rho $ & {\small [0, 1)} & {\small Beta} & {\small 0.8} & {\small 0.1} \\ 
$\phi _{\pi }$ & $%
%TCIMACRO{\U{211d} }%
%BeginExpansion
\mathbb{R}
%EndExpansion
^{+}$ & {\small Gamma} & {\small 1} & {\small 0.5} \\ 
$\phi _{y}$ & $%
%TCIMACRO{\U{211d} }%
%BeginExpansion
\mathbb{R}
%EndExpansion
^{+}$ & {\small Gamma} & {\small 0.1} & {\small 0.25} \\ 
$\rho _{R}$ & {\small [0, 1)} & {\small Beta} & {\small 0.25} & {\small 0.1}
\\ 
$\rho _{y}$ & {\small [0, 1)} & {\small Beta} & {\small 0.25} & {\small 0.1}
\\ \hline\hline
\end{tabular}

\bigskip 

\bigskip 

\bigskip 

\begin{tabular}{||c|c|c||}
\hline\hline
\multicolumn{3}{||l||}{\textbf{Table 2\ \ Bayesian estimates the New
Keynesian\ model's }} \\ 
\multicolumn{3}{||l||}{\textbf{structural parameters: median and 90\%
coverage percentiles}} \\ \hline
&  & After the Volcker \\ 
{\small Parameter} & Great Inflation & stabilisation \\ \hline
$\sigma _{R}^{2}$ & 2.2349 [1.6334 3.1215] & 0.2931 [0.2217 0.3998] \\ 
$\sigma _{\pi }^{2}$ & 2.1693 [1.6085 3.0791] & 1.0515 [0.8039 1.3877] \\ 
$\sigma _{y}^{2}$ & 0.3925 [0.2649 0.6265] & 0.2054 [0.1321 0.3054] \\ 
$\sigma _{s}^{2}$ & 0.1413 [0.0666 0.3925] & --- \\ 
$\theta $-1 & 25.3491 [14.4340 36.1992] & 16.7009 [9.8853 25.0545] \\ 
$\alpha $ & 0.6314 [0.6009 0.6579] & 0.6281 [0.5979 0.6561] \\ 
$\epsilon $ & 0.5955 [0.4982 0.6770] & 0.0922 [0.0136 0.2054] \\ 
$\sigma $ & 6.9248 [4.7527 9.7772] & 2.7398 [1.9696 3.8269] \\ 
$\delta $ & 0.6050 [0.4678 0.7634] & 0.1220 [0.0153 0.3059] \\ 
$\rho $ & 0.6772 [0.5966 0.7459] & 0.8054 [0.7595 0.8405] \\ 
$\phi _{\pi }$ & 1.9782 [1.4931 2.4390] & 2.1690 [1.7494 2.7049] \\ 
$\phi _{y}$ & 1.2221 [0.7075 1.8749] & 0.6813 [0.0552 1.4674] \\ 
$\rho _{R}$ & 0.3232 [0.1742 0.4595] & 0.4451 [0.3349 0.5415] \\ 
$\rho _{y}$ & 0.5115 [0.3732 0.6217] & 0.7799 [0.7178 0.8268] \\ \hline\hline
\end{tabular}%
\bigskip 

\bigskip 

\bigskip 

\bigskip 

\bigskip 

\end{document}

[sitex]

tcitex.tex

(16.9 KiB) Downloaded 1164 times

Hello,
The attached file should format using your system. You should only make changes to the file using Scientific WorkPlace or Scientific Word, and you should save modified files using Save As Portable LaaTeX file.

Regards,
Tom Price

[lucianojcampos]

Hi Tom, thanks so much for your help. The file you´ve sent me works perfectly well.

Best,