I love finding a genuinely new sentiment series to add to the Sentiment Matters toolbox. So I was very happy to discover that Matteo Iacoviello and Jonathan Tong have built an AI-based upgrade to the Geopolitical Risk indices — and, even better, they’ve added a sub-index for oil/energy disruptions.
Why it’s exciting:
- It’s not just keyword counting. The new approach uses AI to read the articles (semantic understanding) and assign a continuous geopolitical-risk score.
- It breaks out oil/energy disruption risk driven by geopolitics — which is… timely.
The good news
It behaves exactly like a useful contrarian sentiment indicator should behave — for both oil and the Energy equity sector.
In my backtests:
- When GPR Oil > 270, oil has on average fallen ~6% over the next 12 months (vs +8% in a typical 12-month window), with below-average returns 77% of the time.
- For the Energy sector, the pattern is similar: average ~10% underperformance over the next year, underperforming 76% of the time.