SIMUL 3.2: An Econometric Tool for Multidimensional Modelling

• Authors: Rodolphe Buda
• Venue: Computational Economics
• Year: 2013

Initially developed in the context of $${\tt REGILINK}$$ project, $${\tt SIMUL 3.2}$$ econometric software is able to estimate and to run large-scale dynamic multi-regional, multi-sectoral models. The package includes a data bank management module, $${\tt GEBANK}$$ which performs the usual data import/export functions, and transformations (especially the RAS and the aggregation one), a graphic module, $${\tt GRAPHE}$$ , a cartographic module, $${\tt GEOGRA}$$ for a "typical use". For an "atypical use" the package includes $${\tt CHRONO}$$ to help for the WDC (Working Days Correction) estimation and $${\tt GNOMBR}$$ to replace the floating point arithmetic by a multi-precision one in a program. Although the current package includes a basic estimation's (OLS) and solving's (Gauss---Seidel) algorithms, it allows user to implement the equations in their reduced form $${Y_{r,b}=X_{r,b} + \varepsilon}$$ and to use alternative econometric equations. $${\tt SIMUL}$$ provides results and reports documentation in ASCII and $${\hbox{\LaTeX}}$$ formats. The next releases of $${\tt SIMUL}$$ should improve the OLS procedure according to the Wilkinson's criteria, include Hildreth---Lu's algorithm and comparative statics option. Later, the package should allow other models implementations (Input---Output, VAR etc.). Even if it's probably outclassed by the major softwares in terms of design and statistic tests sets, $${\tt SIMUL}$$ provides freely basic evolutive tools to estimate and run easily and safety some large scale multi-sectoral, multi-regional, econometric models.