What are competing software packages

The software JMulTi

By Alexander Benkwitz

Abstract

The dissertation develops and examines methods for the analysis of dynamic multi-equation models (VAR models). First a general concept for the integration of statistical procedures into a menu-driven software is developed. The resulting Java library consists of configurable user interface components and functions that enable communication with the statistical software package GAUSS. This library is the basis for the JMulTi software, a menu-driven program for analyzing univariate and multivariate time series. The use of JMulTi in the analysis of VAR models is then documented. For this purpose, unrestricted and restricted VAR models are estimated for the monetary sector in Germany and different bootstrap confidence intervals for impulse responses are calculated and compared. These intervals are the subject of a final and detailed analysis. It is investigated whether the bootstrap method used in JMulTi (and further suggestions such as subsampling) are able to overcome the possible inconsistency of the standard asymptotic method when calculating confidence intervals for impulse responses. The thesis develops and examines tools for the analysis of dynamic multi-equation models (VAR models). First, a general concept for the integration of statistical procedures into a menu controlled software is developed. The resulting Java library consists of configurable graphical user interface components and functions, which allow communication to the statistical software package GAUSS. This library is the basis for the software JMulTi, a menu-driven program for analyzing univariate and multivariate time series. The use of JMulTi for analyzing VAR models is documented next. Unrestricted and restricted VAR models for the monetary sector of Germany are estimated and different bootstrap confidence intervals for impulse responses are computed and compared. These intervals are subject of a concluding and detailed analysis. It is examined whether the bootstrap methods used in JMulTi (and further suggestions, e.g. the subsampling) are able to overcome the possible inconsistency of the standard asymptotic method when computing confidence intervals for impulse responses. A Monte Carlo study illustrates the performance of the examined methods

Topics: VAR Analysis, Bootstrap, JMulTi, Java, VAR Analysis, Bootstrap, JMulTi, Java, 330 Economy, 17 Economy, QH 237, ddc: 330
Publisher: Humboldt University of Berlin, Faculty of Economics
DOI identifier: 10.18452 / 14861
OAI identifier: oai: edoc.hu-berlin.de: 18452/15513
Download PDF:
Sorry, we are unable to provide the full text but you may find it at the following location (s):
  • http: //dx.doi.org/10.18452/148 ... (external link)
  • http: //edoc.hu-berlin.de/18452 ... (external link)