A1498
Title: Media tone in commodity markets
Authors: Ana-Maria Fuertes - City University London (United Kingdom)
Nan Zhao - Cass Business School, City, University of London (United Kingdom) [presenting]
Abstract: Media tone measures, as a proxy for sentiment, are constructed for a cross-section of commodities in the main sectors-energy, agriculture, and metals-based on the aggregate textual tone retrieved from news in Bloomberg terminals. Building on the long-short portfolio with equal-weighted constituents sorted on their hedging pressure characteristic as a baseline, we design an alternative hedging pressure sentiment-weighted portfolio that weighs the constituents according to their media tone over the preceding week. A constituent of the long (short) leg of the baseline portfolio is assigned a smaller (larger) weight if its media tone is net positive (negative), since this favourable (adverse) sentiment is likely to be already reflected in the current futures price pushing it upwards (downwards). Thus, the expected return would be lower (higher) than it would have been anticipated according to the hedging pressure characteristic alone. We find that the media tone corrected portfolio could earn a significantly higher return than the benchmark. Then, we construct a media tone-based sentiment index per commodity. Using it as a trading signal in an out-of-sample portfolio analysis, we find a sizeable and significant sentiment premium after controlling for traditional commodity risk factors. The findings suggest that sentiment can predict the cross-section commodity futures return.
