Predictive power of Google searches on commodity futures: Evidence from US and UK markets

Abstract

This paper explores the effect of investor attention on trading volumes, open interests, returns, and price volatility of commodity futures in the US and UK markets. As a proxy for investor attention, the study uses abnormal search volume indexes (ASVIs) of keywords related to commodities. The study focuses on the immediate effects of investor attention within a time window of 0 to 2 days. The robustness of the results is ensured by separately examining the effect for different commodities, time periods, and variations of the ASVI variables.

Daily financial data for commodity futures contracts is gathered from Refinitiv Datastream for the period from 2013 to 2022. The final sample consists of 22 commodities from US and UK exchanges across three categories: metals, energy, and agriculture. Search volume data for the corresponding commodities is gathered from Google Trends for the commodity names, and the names combined with trading-related identifiers.

The findings suggest that for most commodities, an abnormal increase in search volumes has a significant positive impact on both trading volume and price volatility. For commodities significant at the 1% level, a one-standard-deviation increase in abnormal search volume for a commodity name results in an average increase of 111 bps in trading volume and 1359 bps in the absolute value of the price change on the same day when search volumes peak. However, there is no robust evidence supporting a significant positive relationship between ASVI and returns or open interest. Overall, the results indicate that investor attention influences the dynamics of the commodity futures market, although the effects differ from those observed in equity markets. This discrepancy likely reflects the dominant market share of institutional investors and the less restrictive short-selling constraints in the commodity futures market.