My friend's daughter doing her Ph.
D in Down Under asked me to advise her on software to use for Input-Output
analysis. Sometime before, I was lucky to have looked for such software here in
Myanmar at the request of my younger friends from the academia. With 0% of I-O
analysis knowledge and less than 1% of knowledge for the R statistical
environment I found out about the rwiot
software package developed by UNIDO, thanks to Google. It was destined for the
main repository of R: cran or the comprehensive R archive network. But it wasn't
available on cran that time, and
still not yet available on it. However, the developers directed me to the
download site and anyone can download the two packages, rwiot and rwiotData, as
Windows binaries, from the following links:
So I immediately replied to her about rwiot. Afterwards I noticed that, in a hurry, I missed the point
that she was thinking about doing the I-O analysis via the GEMPACK software. I
looked it up and found that it is for making general equilibrium economic analysis.
With a slant for R, I looked up for general equilibrium modeling in R and I was led
to the gEcon package: http://gecon.r-forge.r-project.org/
and its description reads:
About gEcon
gEcon is a framework for developing and solving large scale dynamic
(stochastic) & static general equilibrium models. It consists of
model description language and an interface with a set of solvers
in R. It was developed at the Department for Strategic Analyses
at the Chancellery of the Prime Minister
of the Republic of Poland as a part of a project
aiming at construction of large scale DSGE & CGE models of
the Polish economy.
Publicly
available toolboxes used in RBC/DSGE modelling require users to derive the
first order conditions (FOCs) and linearisation equations by pen
& paper (e.g. Uhlig’s tool-kit)
or at least require manual derivation of the FOCs (e.g. Dynare).
Derivation of FOCs is also required by GAMS and GEMPACK — probably
the two most popular frameworks used in CGE modelling. Owing to
the development of an algorithm for automatic derivation
of first order conditions and implementation of a comprehensive
symbolic library, gEcon allows users to describe their
models in terms of optimisation problems of agents.
To authors' best knowledge there is no other publicly available
framework for writing and solving DSGE & CGE models in this
natural way. Writing models in terms of optimisation problems instead of
the FOCs is far more natural to an economist, takes off
the burden of tedious differentiation, and reduces the risk
of making a mistake. gEcon allows users to focus
on economic aspects of the model and makes it possible
to design large-scale (100+ variables) models. To this end, gEcon provides template mechanism (similar
to those found in CGE modelling packages), which allows to declare similar
agents (differentiated by parameters only) in a single block. Additionally, gEcon can automatically produce a draft
of LaTeX documentation for a model.
The model description
language is simple and intuitive. Given optimisation problems, constraints
and identities, computer derives the FOCs, steady state equations,
and linearisation matrices automatically. Numerical solvers can be then
employed to determine the steady state and approximate equilibrium
laws of motion around it.
She could download it from: http://gecon.r-forge.r-project.org/download.html.
If gEcon has
capability comparable to GEMPACK I felt it would be good to select the gEcon (i) as GEMPACK is not free, and (ii)
even in case that funds are available for it from the university. Notwithstanding my zero knowledge of general equilibrium modeling, I am certainly
impressed with gEcon when the
introduction says:
To authors' best knowledge there is
no other publicly available framework for writing and solving DSGE
& CGE models in this natural way. Writing models in terms
of optimisation problems instead of the FOCs is far more natural
to an economist, takes off the burden of tedious
differentiation, and reduces the risk of making a mistake. gEcon allows
users to focus on economic aspects of the model and makes
it possible to design large-scale (100+ variables) models.
So by choosing gEcon,
when she comes back to her own country she could freely share the gEcon software as well as her knowledge
with the local researchers as much as she would like.
The difference between GEMPACK and gEcon, I guess, will be that the former would be a stand-alone package (?) while the latter
runs on R. That means she would need to learn the basics of R before she could
use gEcon. However, that shouldn't be
too difficult. I think her university would have some course on R so she could
have started running gEcon in no
time, or she could try any number of tutorials available on the Web. Anyway,
she and her supervisors will be ones to decide if that would be GEMPACK or any
available alternative they would actually use.
On the other hand, I guess that the establishment and the
academia are basically suspicious of free and open source software, for
example, and have a general distrust of something that is free—in the sense
that it must somehow be inferior, seen in such examples as Wikipedia vs.
Britannica, crowd vs. experts, collaborative vs. hierarchical and so on.
One difference is that gEcon
is hardly anything of some free-wheeling software. It is a tool developed specifically
for use by the Polish government and so it is worthy of being taken seriously
by anyone. Therefore I should be more than happy if this message got through to
any interested Myanmar researcher.
.jpg)
No comments:
Post a Comment