Investor attention and stock market behaviour in Europe: Evidence from Google search data

Abstract

This thesis contributes to the growing literature on investor attention by reinforcing the usefulness of internet search data as a powerful tool for both forecasting and nowcasting stock market performance. Specifically, it presents an empirical analysis that applies both nowcasting and forecasting methodologies using aggregated Google search data through the Google Search Volume Index (GSVI) across European stock markets. To provide a solid foundation for the research, the thesis includes a comprehensive literature review and develops a theoretical framework that incorporates the primary theories on investor behaviour and informational demand.

The GSVI reflects the real-time demand for information by investors and, consequently, their intentions and behavioural responses to market stimuli. The central objective of this thesis is to assess the capacity of the GSVI to anticipate and predict movements in the European stock market. To achieve this, three pairs of hypotheses are proposed to analyze, through panel data regression models, the capacity to estimate and forecast three key financial metrics: stock return, volatility, and trading volume.

The empirical analysis is conducted using panel data from 52 publicly traded companies listed in the European stock market and included in the EUROSTOXX 50 index over the period 2014–2024. This cross-country sample enhances the relevance of the findings by offering a broader, more generalizable perspective beyond country-specific stock indices. Three distinct regression models are estimated to determine the influence of the GSVI on returns, volatility, and trading volume.

The results from the forecasting models show that aggregated online search behaviour holds significant predictive power, particularly over trading volume and volatility. Moreover, this predictive capacity increases as the forecasting horizon extends. Thus, the thesis concludes that early signals of investor attention reinforce their value in financial forecasting and, consequently, in investment analysis and decision-making.

Building on these results, the thesis validates the GSVI as a reliable proxy for investor attention in European stock markets, supporting the integration nontraditional data in behavioural finance models. In particular, concerning nowcasting, while its capacity to reflect an immediate perception of financial market conditions is limited, it proves valuable as a complementary indicator for anticipating very short-term changes in returns, volatility, and trading volume.

Moreover, the thesis underscores the importance of decisions related to the construction of the aggregated search index. It offers relevant insights, showing that alternative methods of building the GSVI are aligned with the financial variables analysed and the different forecasting horizons. These conclusions have practical implications for portfolio risk management, sentiment-driven trading strategies, and the development of more effective market signals. Finally, the thesis outlines future research directions to further enhance the use of digital attention metrics in financial markets.