PARADIGM SHIFT IN MAIN MACROECONOMIC VARIABLES’ ANALYSIS

Abstract

Bayesian Vector Autoregressive (BVAR) models are mostly used in computational analysis of
macroeconomic variables of a nation. Certain issues had led to imprecise estimation by BVAR
Model. Recently, it has been discovered that BVAR exhibit some statistical imprecision and
inaccuracy in evaluating and forecasting macroeconomic variables. Because of this inconsistency
with VAR models, the focus has gradually shifted to models that can account for the comovement
among macroeconomic variables. This shift is occasioned by the limitations
associated with BVAR and its inability to elucidate information relating to the integration and
cointegration among the macroeconomic variables. Also, the predictive accuracy of BVAR
model has some blemish owing to the fact that it lacks the estimation capacity to elucidate vital
information about the co-movement among macroeconomic variables of a developing economy
like Nigeria. The research utilizes secondary data of the selected macroeconomic variables which
was obtained from the statistical bulletin of the Central Bank of Nigeria (CBN) ranging from
1986 to 2019. Bayesian Vector Error Correction Model (BVECM) was used in the analysis of
the macroeconomic variables. It was discovered among others that BVECM has more analytical
features than BVAR. Therefore, in forecasting and evaluating main macroeconomic variables of
a nation, BVECM should be considered because it possesses a better predictive ability than the
BVAR model since it predicts the direction of change in the chosen macroeconomic variables.