What the study found
The study found that the time-varying Scalar Component VARMA (TV-SCVARMA) model produced the lowest root mean squared error and stable forecasts for Nigerian macroeconomic data. It also found that classical VARMA performed poorly.
Why the authors say this matters
The authors conclude that time-varying models are better suited for forecasting in unstable economies. They recommend TV-SCVARMA for macroeconomic policy analysis and suggest future work could extend the model to stochastic volatility or Bayesian estimation for extreme-shock environments.
What the researchers tested
The researchers developed and estimated a TV-SCVARMA model, which is a state-space model that lets parameters change over time while keeping the parsimony of VAR and retaining moving-average dynamics. They used quarterly Nigerian data from 2010Q1 to 2024Q1 on real GDP growth, inflation, money supply (M1 and M2), and exchange rate, and estimated the model with a Kalman filter-based maximum likelihood approach.
What worked and what didn't
The TV-SCVARMA model showed the lowest RMSE, rapid convergence, and stable forecasts. Classical VARMA performed poorly, and the abstract says traditional VAR and VARMA approaches are unreliable when parameters shift over time.
What to keep in mind
The abstract describes macroeconomic instability in Nigeria and evaluates the model on quarterly data from 2010Q1 to 2024Q1, so the findings are scoped to that setting. Limitations beyond the noted identification and small-sample problems of classical VARMA are not described in the available summary.
Key points
- TV-SCVARMA had the lowest RMSE among the models tested.
- The model produced stable forecasts and converged quickly.
- Classical VARMA performed poorly in the comparison.
- The data covered Nigeria from 2010Q1 to 2024Q1 and included GDP growth, inflation, money supply, and exchange rate.
- The authors recommend time-varying models for forecasting in unstable economies.
Disclosure
- Research title:
- TV-SCVARMA outperforms classical VARMA in Nigeria
- Authors:
- M. O. Osolo, C. N. Okoli, M. S. Laisin
- Institutions:
- Chukwuemeka Odumegwu Ojukwu University
- Publication date:
- 2026-02-24
- OpenAlex record:
- View
Get the weekly research newsletter
Stay current with peer-reviewed research without reading academic papers — one filtered digest, every Friday.