Markov-Switching Vector Autoregressive Modelling (Intercept Adjusted); Application to International Trade and Macroeconomic Stability in Nigeria (2000M1–2019M6)

Abstract

Economic relationships are often modelled without consideration of a possible regime switch, the transmission from one regime to another and the duration of stay in a particular regime which are not captured by linear models. This study aimed to model and estimate the interdependence existing among Nigeria’s International Trade and Macroeconomic Stability. Specifically, this study sought to estimate and compare the estimated Models, select the best Model and determine the probabilities of stay, the expected duration of stay in a particular regime. The study adopted a quasi-experimental design. Time series data on the study variables from January 2000 to June 2019 were obtained from the Statistical Bulletin of the Central Bank of Nigeria. Models were specified accordingly, the statistical analyses were carried out using the Markov Switching Intercept Vector Autoregressive Models, the pre and post-diagnostic tests were also conducted. The unit root test results showed I (1). VAR lag length selection criteria choose lag 2. The MS-VAR analysis identified two regimes (expansion and contraction), the information criteria selected the Markov-Switching Intercept Autoregressive Heteroschedastic 2 Variance Auto-regression 2 [MSIARH (2) - VAR (2)]. The MS-VAR results in regime 1 showed that lags 1 and 2 of total export significantly affected total export and total import, Lags 1 and 2 of total import had significant effects on exchange rate while lags 1 of exchange rate and lags 1 and 2 of exchange rate had significant effects on inflation rate. In Regime 2, lag 1 of total export and lag 2 of exchange rate had significant effects on total export. Only lag 2 of inflation rate had significant effects on exchange rate while lag 2 of total export and lags 1 and 2 of exchange rate had significant effects on the inflation rate. The results also showed an 89% probability of staying in regime 1 for a duration of 8 months 8 days and 57% probability of staying in regime 2 for 2 months 10 days. It was concluded that the MSIARH (2) - VAR (2). It was recommended that the right-hand side variables should be tested for endogeneity before concluding on single or system equation. It was also recommended that the possibility of regimes should be verified before concluding on linear or nonlinear models.