The first statistical software packages developed for personal computers landed midway through the Eighties. Thirty years on, the best programs offer countless possibilities for financial econometricians, at a high cost. There are, however, open-source alternatives.
10 december 2015
Data, data, and more data. Financial econometrics depend on agile and effective ways of working with data. In the last quarter-century the total computerisation of the financial world has allowed for the creation of the ideal tools for making progress in this field.
Many of these programs – a majority – require a private licence to use, with further renewal fees, even for universities and specialised schools. But in the last decade the number of open-source options that imitate these programs at no cost have multiplied. They also come with an advantage, the freedom for users to modify, enhance and add-on to the software (through patches). Here are some of the most useful software packages for financial econometricians:
R is in itself a programming language, in the same vein as Java, C or C++. As a free implementation of S programming language (with certain modifications that allow for the easier application of specific mathematical functions), it provides a vast array of options for carrying out operations with matrices – the ideal atmosphere for producing econometric packages. A command-line user interface is provided, but there are also a number of installable GUIs (graphical user-interface) available that aid in levelling out the steep learning curve. An extensive list of complementary packages specialising exclusively in financial econometrics is also available.
The acronym for GNU Regression, Econometrics and Time series Library, Gretl is a software package specialising in econometric analysis, accompanied by a large catalogue of function packages in constant development and cross-software support (both free and licensed). It is much more user-friendly than R for beginners, although for certain advanced econometric analysis functions it may come up short. By combining the graphical interface with the manual input of code, Gretl becomes a very adequate program. Curious fact: it’s distributed in over twenty languages, including Basque, Galician and Catalan.
SPSS, now part of the IBM family, is one of the oldest statistical packages in the world, in continuous development since 1968. It was also one of the first to reach home computers, in 1984. Scholars have benefitted from this software for nearly half a century. However, it also carries two major drawbacks (lack of compatibility with Unix operating systems and pricing – the cheapest licence is still over 1000€) that created an opening for PSPP. It can be relied on for basic to intermediate statistical analyses and it supports an impressive amount of variables and cases (over a billion of each), but it is outdated and lacks many advanced options, graphical creation in particular.
Many other software packages are available for free, with (more or less) adequate uses for financial econometrics: Octave or Scilab are particularly handy for matrix analysis (and they carry similar attributes to licensed software MatLab), JMulti specialises in time-series, and other programs can be approached for specific purposes, such as typesetting. Every one of them free of charge, free to use, and often just as functional as their commercial counterparts. So yes, there is a world beyond EViews.
