COINT 2.0: Co-integrated Systems

Sam Ouliaris and Peter C.B. Phillips

The following product is developed by Sam Ouliaris and Peter C.B. Phillips, third party developers, for use with GAUSS. Technical support is provided directly through the developers.

COINT 2.0 - Co-Integrated Systems

A suite of econometric software for GAUSS users with a special focus on nonstationary time series, unit roots, cointegration and modern model selection methods for economists, econometricians, statisticians, engineers, forecasters and other users of time series methods.

Whether you are an economist doing empirical time series research, an econometrician in a forecasting unit, a professor teaching econometrics or a graduate student of economics or statistics, you need access to the latest regression methods for stationary and nonstationary time series.

COINT gives GAUSS users a huge library of scientific procedures for time series regression and model selection. Included are the latest techniques for unit root testing, cointegrating regression estimation, ARMA and VAR modeling with some unit roots, GMM and GIVE estimation with nonstationary data, and Bayesian as well as classical statistical methods for detecting unit roots and cointegration in economic time series.

COINT will enhance your research and teaching by giving you access to state-of-the-art times series methods and econometric techniques. Be more productive in GAUSS, work with the latest nonstationary regression methods and give presentations that utilize the latest features of GAUSS publication quality graphics.

Platforms: Windows, LINUX, UNIX

Requirements: GAUSS version 3.2 and above.
Sam Ouliaris and Peter C.B. Phillips
Research Department, HQ 9-711B
International Monetary Fund
700 19th Street, NW
Washington, DC 20431
Phone: Sam: 202-623-8009   Peter: 203-432-3695
Fax: Sam: 202-589-8009   Peter: 203-432-3703
Email: souliaris@imf.org
peter.phillips@yale.edu

COINT 2.0 gives you:

Unit Root Tests

Have a wide range of procedures at your fingertips to test for the presence of a unit root. Use the latest data-based tests that employ model selection and kernel estimation with automatic bandwidth selectors. Test your data for stationarity as well as nonstationarity.

Cointegration Tests

Test for cointegration and find the dimension of the cointegration space using data-driven residual based tests and likelihood ratio tests.

Tabulated Critical Values

Have at your disposal a complete set of tabulated critical values for unit root and cointegration tests. COINT has an automated search facility that delivers critical values whenever test statistics are computed.

Cointegrating Regression

Choose a routine for estimating the parameters of a cointegrated system. COINT has a large selection of methods: FM-OLS and its latest enhancements including FM-GMM, FM-GIVE and FM-VAR; reduced rank regression methods; canonical cointegrating regression; spectral regression; and structural stability tests for cointegrating regression.

Bayesian Unit Root Analysis

Do a Bayesian analysis of nonstationarity for your time series and cointegrating regression residuals. COINT gives you graphical procedures to plot marginal posterior densities and calculates posterior probabilities of nonstationarity.

ARMA Model Selection and Estimation

Estimate ARMA models by recursive techniques that include automated order selection procedures. Choose a model selection method like AIC, BIC or PIC, find a suitable model for your data and use graphical displays with built-in unit root tests in your evaluation.

Kernel Estimation

Access a full library of kernel estimation routines for the estimation of spectra, long run variances, and one-sided long run covariances. Data-driven bandwidth methods are available as well as the latest AR- and ARMA-prefiltered kernel procedures.

COINT is supplied with a complete reference manual for the use of all of its procedures, a bibliography to the literature, and full instructions for set-up and installation with GAUSS. COINT is supplied in GAUSS source code so that, as a user, you have access to the code for your own personal use in teaching and research.