Assessing the Predictive Might of Vector Autoregressive and Autoregressive Distributed Lag Models: Evidence from Nigeria
Abstract
This paper aim to analyze the short-run relationships between key macroeconomic variables in Nigeria, while determining whether Autoregressive Distributed Lag or Vector Autoregressive model provides superior predictive accuracy. The data set encompasses Inflation rate, Exchange rate, Imports and Crude oil price between January, 1995 to March, 2024 from central Bank of Nigeria data website. The study adopted ARDL Bounds test and VAR Johansen cointegration test. Results show that from the VAR estimates, all four variables are highly persistent and primarily driven by their own historical movement. Also the ARDL model comparatively possesses higher predictive power for this data set. In terms of model nuances, the VAR model demonstrates advantage in minimizing the Mean Absolute Percentage Error for exchange rate. Exchange rate and import level significantly influence crude oil price underscoring the critical role of lag effects in macroeconomic estimation. It is recommended that Policymakers should therefore closely monitor and manage import and exchange rate policies. Key Words: Autoregressive, Vector, Predictive, Lag, Model, Macroeconomic
Keywords
References
More Articles from RESEARCH JOURNAL OF PURE SCIENCE AND TECHNOLOGY
Author: Ngonadi Uchechi, Michael Ominyi, Ngozi Samuel Uzougbo, Blessing Chika Jones,, USA
Author: Awonowo Olusegun Oriyomi,, Anrinnle Qowiu Olaoluwa, Lawal Ahmed Oladimeji, Babayemi Temitayo Matthew, Azubuike Marvelous Onyedikachukwu, Olomu Ayomide Babatunde, Omitogun Ayomide Elijah
Author: Olatunde Adesoji Aladegba, Moshood Ayinde
Author: Olatunde Adesoji Aladegba, Moshood Ayinde
Author: A. Musa, B. B. M. Dewu, A. A. Gado, and I. U. Mohammed, Corresponding E-mail
