Deep-π*: a neural network estimate of trend inflation
The Neural Herald summary, AI-generated
Notice on the use of AI. In compliance with Article 50(4) of Regulation (EU) 2024/1689 on Artificial Intelligence (AI Act), we disclose that the summaries, explanations and translations published by The Neural Herald are generated with artificial intelligence and published without prior human review. The linked official documents belong to each body and were not generated by our AI. Our summaries may contain errors: always check the original document.
Working paper proposing Deep-pi*, a neural network estimate of trend inflation that attributes its movements to macroeconomic indicators.
Key points
- Using US data, employment was the main contributor to the rise in trend inflation after the pandemic; the forward-looking trend rose earlier and peaked 8 to 12 months before
- Both trends ignore most energy and commodity movements
- Better predictive performance than the standard UC-SV model during the inflation surge
Who is affected
Inflation research
Plain-language version
An ECB study (the ECB is the euro area's central bank) uses neural networks, a type of AI, to estimate trend inflation, the underlying direction of prices without short-term swings. Using US data, it finds jobs were the biggest driver after the pandemic, and the model predicted the price surge better than the standard method. Measuring it well helps get interest rate decisions right.
Read the original document →
Prompt for your AICopies a ready-to-paste prompt for ChatGPT, Claude or Gemini: it asks the model to read the official document and not to make anything up.
Summary produced with artificial intelligence from the official publication. It may contain errors: always check the original document.
More from ECB
- 13 November 2026Isabel Schnabel: Made in Europe
- 12 November 2026Economic Bulletin Issue 7, 2026
- 12 November 2026Artificial intelligence and European firms: the cost of market-based financing
- 12 November 2026What share of public services employment is contributed by the private sector?
- 12 November 2026Narratives and risks: a quantitative toolkit to evaluate scenarios applied to the Middle East conflict
- 12 November 2026Reading the room: what Reddit tells us about retail investor sentiment