How Does the Predictive Relationship Between Google Trends Search Volume Index and CAC40 Trading Volume Differ Between Market Stress and Calm Periods?

<span>Investor attention measured via Google Trends Search Volume Index (SVI) has been found to predict stock market activity, but no prior studies have compared prediction power across SVI types. Prospect theory’s loss aversion suggests that attention - trading volume relationship strengthens during market stress, while Peng et al. (2007) prediction suggests a two-phase contemporaneous spike and a reversion period. Three daily SVI series (Simple, Composite, Negative) measuring investor attention were analyzed against CAC40 trading volume data for 1834 trading days, spanning 2019</span><span>–</span><span>2026, and classified “calm” or “stress.” Granger causality, moderated regression, and a structural vector autoregressive model (SVAR), were used to analyze both phases, with extensive robustness checks. Composite SVI was found to predict volume only during stress, consistent with Peng et al. (2007) prediction. SVAR was significant across all types and regimes, with a 5-day decay period, following phase 2. This study provides first direct confirmation of Peng et al. (2007) prediction and demonstrates that single-method SVI analyses may produce misleading results.&nbsp;</span>