Supply and Demand Side Sentiment Drives Crude Oil Prices

Brandt & Gao (Duke University / University of Luxembourg) use RavenPack's Big Data analytics to predict oil prices.

Oil prices are notoriously hard to explain and predict. The academic literature has suggested anything from real prices of oil following a random walk without drift, to the explanatory power of oil pricing factors to vary over time and with different importance at different time horizons. The ever changing nature of this predictive relationship contributes to the difficulty of forecasting oil prices. Furthermore, oil prices are not only related to economic fundamentals but also to geopolitical events that are much harder to quantify. However, with the emergence of big data analytics, datasets have become more easily available that quantify both macroeconomic and geopolitical events - offering new avenues to explore for “alpha capture”.

In a recent study, Brandt (Duke University) & Gao (University of Luxembourg) used RavenPack data to compare and contrast the importance of macroeconomic and geopolitical information for crude oil.

Focusing on 29 news categories that were considered relevant for the oil market, including terrorism, war & conflict, civil unrest, natural disasters etc.; they calculate an aggregated daily news index. As a unique feature of the index, they differentiate between news on oil-producing versus oil-consuming countries, since a contrasting impact on oil prices is often observed. For example, civil unrest in oil-consuming countries is generally expected to reduce oil supply and hence lead to a price increase, while unrest in oil-consuming countries is typically associated with contraction in the economy, and hence also in the oil price. Figure 1 and 2, plots the Geopolitical and Economic sentiment indexes that was used in the study against the oil price.

alt text

alt text

Below, is an overview of the key findings of Brandt & Gao’s research:

  • Using sentiment scores for a broad set of global news of different types, they find that news related to macro fundamentals has an impact on the oil price in the short run and significantly predict oil returns in the long run.
  • Geopolitical news has a much stronger immediate impact but exhibit no predictability beyond the intraday horizon.
  • Moreover, geopolitical news generates more uncertainty and greater trading volume, consistent with a disagreement explanation, while macroeconomic news reduces informational asymmetry and is associated with subsequent lower trading volume.
  • Finally, they found that news sentiment contains more information about future expectations than about future realizations of economic data.

All in all, considering news analytics as part of an oil price prediction model, we’re able to include information which is otherwise hard to capture. This allows for new avenues to explore when it comes to alpha capture.

Discover the RavenPack Platform

We offer more than 90 dashboards, built with over 150 curated datasets and powerful visualisation tools.