<![CDATA[This research studies a Bayesian Vector Autoregressive (BVAR) model to analyze the dynamic interactions among the rupiah exchange rate, exports, imports, gold futures prices, and inflation in Indonesia during the 2015-2024 period. The BVAR method was chosen to overcome the limitations of conventional VAR models on overparameterization problem by utilizing hierarchical Minnesota priors and Markov-Chain Monte Carlo (MCMC) estimation. Data were stationary through first order differencing and normalized using z-score. Lag selection based on the Akaike Information Criterion (AIC) showed that lag 6 is optimal. Model evaluation using Mean Absolute Percentage Error (MAPE) shows good overall model performance on training data, especially on the gold price variable (MAPE 10,09%) and inflation (MAPE 3,74%). On test data, the model struggles to perform well on prediction due to the high uncertainty of the test data period. Impulse Response Function (IRF) analysis is used to reveal short-term responses between variables, such as the effect of exchange rate depreciation on inflation and the impact of export value on a temporary decline in import value. The result highlights the BVAR model’s ability to capture general macroeconomic relationships, especially when many parameters need to be estimated and the available data is limited.]]>
Copyrights © 2026
