Forecasting Macroeconomic Variables and their Effect on Poverty

Abstract

Forecasting macroeconomic variables is crucial to measure dynamic changes during uncertain economic conditions. This study examines and analyzes the appropriate and accurate forecasting model to predict macroeconomic variables in Maluku Province. The main variables used are economic growth, unemployment, inflation, and poverty. The modeling used in this study were Bayesian Vector Autoregressions Model and the Univariate Benchmark Model. The results of this study indicate that the two models have different specifications and forecasting directions. The value of the Univariate Benchmark model’s forecast error size is relatively smaller than that of the Bayesian Vector Autoregressions Model. The results of forecasting macroeconomic variables in Maluku Province have a relatively good level of accuracy and are close to the actual value of the sample period. The Error Correction Model test results show that only the Error Correction Term variable significantly affects the poverty level in the short term. Meanwhile, in the long term, the unemployment rate has a significant effect, and the model used is proven valid. The forecasting results from the model show that the Maluku provincial government must maintain the stability of macroeconomic variables, especially the inflation rate and unemployment
rate, because they tend to increase in the coming year. It can have an impact on reducing people’s purchasing power.