ANALYSE MULTIVARIÉE DES SÉRIES CHRONOLOGIQUES : APPLICATION SUR LE LOGICIEL STATA
Jean-Claude Nkashama Mukenge, Nathan Mbende · HAL (Le Centre pour la Communication Scientifique Directe) · 2025
This paper presents in detail the multivariate analysis of time series with its application on the STATA software. It is subdivided into several of the following points: (1) generalities on time series; (2) Concepts on chronological data and stationarity; (3) causality; (4) cointegration; (5) presentation of the STATA interface; (6) implementation of data to the STATA software; (7) reminder on univariate and bivariate analysis; and (8) choice of the econometric model. Point (8) is the main point of this document. The multivariate analysis of time series makes it possible to explore the relationships between several variables over time. It is essential to understand economic, financial, and social dynamics. The mastery of these techniques in STATA facilitates the practical application of these theoretical concepts. The document takes turns addressing the following models: simple and multiple linear regression, autoregressive vector model (VAR), structural autoregressive vector model (SVAR), error correction model (ECM), error correction vector model (VECM), distributed delay autoregressive model (ARDL) and the chow structural test.