Incertidumbre en el mercado de bonos: una propuesta para identificar sus narrativas con GDELT
Series: Occasional Papers. 2505.
Author: Jéssica Guedes, Diego Torres, Paulino Sánchez-Escribano and José Boyano.
Full document
Abstract
This study analyses the relationship between financial news narratives and volatility in the US government bond market, as measured by the MOVE Index. We leverage a novel dataset, the Global Database of Events, Language, and Tone (GDELT), which provides metadata such as themes and sentiment for online news articles. Using this dataset, we employ large language models (LLMs) to pre-select relevant themes and apply two techniques to identify those most influential on MOVE fluctuations: a LASSO algorithm to pinpoint news themes impacting the index and, to mitigate multicollinearity, a linear regression on pre-identified theme clusters. Both methods are tested on three periods of heightened MOVE Index volatility since 2017. The results show that news narratives influence bond market volatility and that the LASSO algorithm effectively identifies the most impactful narratives. This study provides valuable insights for investors and policymakers by connecting financial news to bond market volatility, paving the way for future research on the impact of news on financial markets.