From d3c44f4ceb7cfda1c52888f8a8eb4ca8ab77bfe1 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?S=C3=A9bastien=20Villemot?= <sebastien.villemot@ens.fr>
Date: Thu, 8 Sep 2011 18:09:21 +0200
Subject: [PATCH] =?UTF-8?q?The=20variance=20prior=20for=20BVAR=20"=C3=A0?=
=?UTF-8?q?=20la=20Sims"=20with=20only=20one=20lag=20is=20inconsistent;=20?=
=?UTF-8?q?the=20solution=20implemented=20consists=20of=20adding=20one=20e?=
=?UTF-8?q?xtra=20observation=20in=20the=20presample=20used=20to=20compute?=
=?UTF-8?q?=20the=20prior;=20as=20a=20consequence,=20the=20numerical=20res?=
=?UTF-8?q?ults=20for=20all=20estimations=20will=20be=20slightly=20differe?=
=?UTF-8?q?nt=20in=20future=20releases=20(thanks=20to=20Marek=20Jaroci?=
=?UTF-8?q?=C5=84ski=20for=20spotting=20this)=20(cherry=20picked=20from=20?=
=?UTF-8?q?commit=20ff7792a7284baa0392eea1d16c4a439c2691e749)?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
---
doc/bvar-a-la-sims.tex | 8 +++++++-
matlab/bvar_toolbox.m | 2 +-
2 files changed, 8 insertions(+), 2 deletions(-)
diff --git a/doc/bvar-a-la-sims.tex b/doc/bvar-a-la-sims.tex
index b176c2199..c69ed695b 100644
--- a/doc/bvar-a-la-sims.tex
+++ b/doc/bvar-a-la-sims.tex
@@ -102,7 +102,13 @@ Before constructing the dummy observations, one needs to choose values for the f
\item $d$: the decay factor for scaling down the coefficients of lagged values. Controlled by option \texttt{bvar\_prior\_decay}, with a default of 0.5
\item $\omega$ controls the tightness for the prior on $\Sigma$. Must be an integer. Controlled by option \texttt{bvar\_prior\_omega}, with a default of 1
\item $\lambda$ and $\mu$: additional tuning parameters, respectively controlled by option \texttt{bvar\_prior\_lambda} (with a default of 5) and option \texttt{bvar\_prior\_mu} (with a default of 2)
-\item based on a short presample $Y^0$ (in Dynare implementation, this presample consists of the $p$ observations used to initialize the VAR), one also calculates $\sigma = std(Y^0)$ and $\bar{y} = mean(Y^0)$
+\item based on a short presample $Y^0$ (in Dynare implementation, this
+ presample consists of the $p$ observations used to initialize the VAR, plus
+ one extra observation at the beginning of the sample\footnote{In Dynare 4.2.1
+ and older versions, only $p$ observations where used. As a consequence the
+ case $p=1$ was buggy, since the standard error of a one observation sample
+ is undefined.}), one also calculates $\sigma = std(Y^0)$ and $\bar{y} =
+ mean(Y^0)$
\end{itemize}
Below is a description of the different dummy observations. For the sake of simplicity, we should assume that $ny = 2$, $nx = 1$ and $p = 3$. The generalization is straigthforward.
diff --git a/matlab/bvar_toolbox.m b/matlab/bvar_toolbox.m
index 917e6100e..eb0a55cca 100644
--- a/matlab/bvar_toolbox.m
+++ b/matlab/bvar_toolbox.m
@@ -98,7 +98,7 @@ mnprior.tight = options_.bvar_prior_tau;
mnprior.decay = options_.bvar_prior_decay;
% Use only initializations lags for the variance prior
-vprior.sig = std(dataset(options_.first_obs+options_.presample-nlags:options_.first_obs+options_.presample-1,:))';
+vprior.sig = std(dataset(options_.first_obs+options_.presample-nlags:options_.first_obs+options_.presample,:))'
vprior.w = options_.bvar_prior_omega;
lambda = options_.bvar_prior_lambda;
--
GitLab