The role of investor sentiment and financial technology systems on asset pricing and market efficiency
Ο ρόλος του συναισθήματος του επενδυτή και των συστημάτων χρηματοοικονομικής τεχνολογίας στην αποτίμηση των περιουσιακών στοιχείων και στην αποτελεσματικότητα των αγορών

Doctoral Thesis
Author
Tsitsiri, Polyxeni G.
Τσιτσίρη, Πολυξένη
Date
2025View/ Open
Abstract
This thesis examines how investor sentiment and FinTech Systems affect asset pricing and market efficiency in the FinTech equity market of G7 countries in the period of November 2000 - November 2024. Motivated by the increasing digitization of financial markets and trading, increases in algorithmic trading, and the behavioral aspects of investor decision making, this research connects behavioral finance with the innovation-based asset pricing. Although sentiment induced mispricing has long dominated the field of behavioral asset pricing, there has been an increasing empirical and theoretical focus on the importance of intangible capital and digital infrastructures in pricing asset returns.
To model this duality, composite indexes are constructed, the Investor Sentiment Index (ISI), input as turnover and price-to-earnings measures, and the Financial Technology Systems Index (FSI), as firm-level capital expenditures and market capitalization, while R&D and patents are excluded because of data availability constraints. The two indices are generated using PCA, standardized and plugged into ten asset pricing models including the CAPM, Fama-French 3-, 4- and 5-factor models, Carhart’s four-factor model (FFC), HML Devil (FAFF), Hou–Xue–Zhang’s q-factor model, and QMJ and BAB and q5 specifications. The empirical estimation is based on panel regressions captured by fixed and random effects as determined by Hausman tests, and robust to heteroskedasticity.
The findings show that FSI is a statistically and economically priced factor in each model and period, consistently increasing model fit and adjusted R². Unlike ISI, it has small and finite time-dependent predictability, and the significance is mainly in the full sample and the early (2000–2012) subperiod, whereas it is absolutely insignificant in the pre- and post-2020 era (pandemic period). These results suggest a structural change in return formation, from behaviorally driven valuations in FinTech’s formative years to innovation driven pricing regimes in the mature, post-pandemic market.
A number of robustness checks are performed to confirm this finding, including subsample analysis (pre/post-2012 and pre/post-2020), alternative definitions of excess returns: that using global risk-free rates (Fama-French, AQR, q-factors), and model specification diagnostics. The thesis also analyzes the market efficiency through the return autocorrelation and the price delay regression. The findings reveal that return predictability and price delay have the strongest significance as weak-form inefficiencies in the early subperiods, which then suddenly drop. ISI is not able to predict a great deal of inefficiencies in any one period, and FSI can predict speed of price adjustment in the latter years and its impact on information speeding up flow.
Theoretically, the thesis has implications for behavioral and innovation-based pricing models, indicating that investor sentiment is episodic and state-dependent, but that FinTech systems are a persistent structural driver of both return generation and efficiency enhancement. Methodologically, it shows how PCA based factors can be incorporated in panel asset pricing models for unbalanced panels. Empirically, it is able to recognize FinTech intensity as a priced factor and as an explanatory factor of informational efficiency for developed equity markets.
These results are of significant consequences for investors, regulators and policy makers. Moreover, they highlight the need to consider tech-specific drivers in valuation, forecasting, and policy design (particularly in industries where digital infrastructures have a transformative impact market structure). Moreover, the empirical evidence for this study could also shed light on the relationship among behavioral biases, technological frontier and asset pricing in the burgeoning digital society.


