Can the sentiment of the official media predict the return volatility of the Chinese crude oil futures?
This study investigates whether the sentiment of Chinese official media towards crude oil influences price volatility of the Chinese crude oil futures (SC). By leveraging textual analysis through Bidirectional Encoder Representations from Transformers (BERT), we quantify the sentiment of oil-related...
Gespeichert in:
Veröffentlicht in: | Energy economics 2024-12, Vol.140, p.107967, Article 107967 |
---|---|
Hauptverfasser: | , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | This study investigates whether the sentiment of Chinese official media towards crude oil influences price volatility of the Chinese crude oil futures (SC). By leveraging textual analysis through Bidirectional Encoder Representations from Transformers (BERT), we quantify the sentiment of oil-related articles published by the primary official media in China. Our main finding, building on both in-sample and out-of-sample analyses, robustly reveals that this sentiment significantly forecasts the one-day-ahead intraday return volatility of SC. Moreover, we extend the analysis to different time horizons (i.e., one-week and one-month-ahead) and find the prominent forecasting power of the official media sentiment as well. We also find that the official media sentiment fails to forecast the price volatility of WTI oil futures, implying that the official media sentiment contains some unique Chinese information. Overall, our study contributes to the existing literature on predicting the return volatility of the Chinese crude oil futures, and offers fresh insights into an essential yet underexplored sentiment, i.e., official media sentiment.
•Construction of the sentiment index for Chinese official media regarding crude oil.•Robust predictor for intraday high-frequency volatility of SC.•Short-, middle-, and long-term in-sample and out-of-sample predictability results.•Novel insights into the underexplored type of sentiment: official media sentiment. |
---|---|
ISSN: | 0140-9883 |
DOI: | 10.1016/j.eneco.2024.107967 |