Multivariate time series modelling of taxes revenue in Nigeria
Tax is a source of revenue or income for a government to achieve a country’s macro economic objectives in the areas of fiscal and monetary policies. However, the effects of tax burden may cause economic recession, financial crises, as well as poor standard of living and economic hardship for the peo...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2022
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/id/eprint/102493/1/BabaGimbaAlhassanPFS2022.pdf.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Tax is a source of revenue or income for a government to achieve a country’s macro economic objectives in the areas of fiscal and monetary policies. However, the effects of tax burden may cause economic recession, financial crises, as well as poor standard of living and economic hardship for the people. In order to tackle these challenges, there is a need for short-term, medium-term, and long-term periods of forecasting models to be developed. In general, financial time series forecasting models are not tax-revenue based-models. Thus, existing models are inadequate to gauge the relationship between tax-based variables that can be particularly volatile. This research bred a model that used data with distinguished variables, obtained from the bulletin of the National Bureau of Statistics and the Central Bank of Nigeria. In this study, Vector Autoregressive model (VAR) and the functional Generalized Autoregression Conditional Heteroscedasticity family (fGARCH) models were combined to consider the behaviour of the financial time series (tax revenue) data. However, because of the high persistence of volatility in the data, the GARCH family model alone is unable to capture the leverage effects in the structural changes of the time series. Hence, the Auto-regression Hidden Markov Model (ARHMM) was proposed to handle this issue. The results show that the VAR with the hybrid of fGARCH models were unable to capture the volatile behaviour of the tax revenue data. On the other hand, the proposed model that used the ARHMM to capture the intensity of volatility persistence performed better. The out-of-sample forecasting accuracy gave less than ten percent of the Mean Absolute Percentage Error (MAPE) for the proposed model. The simulation study has proven that the VARARHMMfGARCH proposed model produced better results as compared to the hybrid of the traditional VAR-fGARCH model. The newly joint VAR-ARHMM-fGARCH model offers an effective forecasting approach for future tax revenue data. |
|---|