Developing a News-Based Sentiment Index for Indonesian Government Bonds Using Enhanced VADER Sentiment Analysis

Market sentiment plays a crucial role in financial markets, influencing equities, cryptocurrencies, and government bonds. However, quantitative measures of investor confidence in Indonesian Government Bonds (SUN) have been predominantly limited to macroeconomic indicators. This study addresses this gap by developing and validating a news-based sentiment index for Indonesian government bonds from 2012 to 2024, analyzing 220,904 articles across five major financial news platforms. Utilizing an enhanced VADER sentiment analysis framework incorporating Indonesian language lexicons, the study finds a statistically significant relationship between news sentiment and bond yield movements, with an ANOVA F-statistic of 9.15 (p-value: 1.09e-04). Quantile regression results indicate sentiment coefficients ranging from +1.5892 at the 0.25 quantile to -1.6759 at the 0.75 quantile, with the strongest predictive capability observed during high-yield periods (p < 0.001). The proposed sentiment index enables real-time monitoring of investor confidence in SUN, offering a valuable tool for policymakers and investors to anticipate yield fluctuations and market stability. Furthermore, this methodology provides a replicable framework for sentiment analysis in emerging market debt instruments, bridging the gap between computational finance and market application.