Do Investors Rely on Robots? Evidence from an Experimental Study

Barbara Alemanni, Andrej Angelovski, Daniela Teresa di Cagno, Arianna Galliera, Nadia Linciano, Francesca Marazzi, Paola Soccorso

Research output: Working paper


Robo advice has moved its first steps in the Anglo-Saxon countries and is now rapidly gaining market share at a global level. The phenomenon fueled a growing and still not conclusive institutional debate about potential benefits and risks to financial consumers, based also on investors’ biases and behaviours that online platforms could trigger to the detriment of robo advisees. The present paper provides some insights into attitudes and behaviours that might prevail in a digital environment among young investors, representing the category of users potentially more involved by the development of the automated advice. In detail, the study investigates whether individuals’ propensity to follow the recommendation received from an advisor changes depending on whether the advisor is a human or a robot. The analysis is based on data collected through an ad hoc developed laboratory experiment run in the Cesare Lab of LUISS University with a sample of 180 students. Students were given an initial monetary endowment and were asked to choose between six different portfolios of financial activities; after being profiled through a questionnaire aimed at eliciting their risk tolerance (Grable and Lytton’s Risk Tolerance Quiz; 2003), they received the advice, either from a human advisor or from a robo advisor (i.e. via a computer platform) depending on the treatment they had randomly assigned before entering the experimental session. Then, they were asked again to choose among the six portfolios in order to capture whether the propensity to follow the recommendation depends on its source (human versus robo). Finally, participants were asked to answer several questions eliciting risk preferences, financial literacy (actual and perceived) and digital literacy, serving as control variables when modelling the probability to follow the advice. Our results show that the probability to follow the advice does not depend on the source of the recommendation but rather on the alignment between the self-directed choice made before receiving the advice and the recommendation subsequently received: the propensity to follow the advisor (either human or robo) increases if the advice confirms individual’s own beliefs about her/his investor profile. This result might be explained by referring to individuals’ attitude towards the so called ‘confirmation bias’. However, when the self-directed choice differs from the recommendation received, participants may be more likely to follow the advice given by a human advisor and less likely to follow the advice formulated by an algorithm. Also the gender of the advisor seems to matter: women tend to follow the advice provided by a female advisor more frequently compared to the case of the recommendation given by a male advisor. This work is part of a wider research on FinTech that CONSOB started in 2016, in collaboration with several Italian universities, with the aim of exploring opportunities and risks for investor protection and the financial system as a whole, related to the application of technological innovation to the provision of financial services. In particular, supplementing Lener, Linciano and Soccorso (2019, edited by) and Caratelli et al. (2019), this document widens the field of investigation by referring to a specific target of the population - the so called millennials and post-millennials – and using complementary and innovative methods. According to an evidence-based approach, insights from the present study may suggest specific investor protection initiatives, also in terms of financial education activities designed for a clearly-identified segment of the population (the so called millennials and post-millennials, in this case). Evidence from the present work might be extended further with respect to the consumers’ perception of the fairness of algorithms used to provide financial services, the cognitive heuristics and biases underlying de
Original languageEnglish
Number of pages61
Publication statusPublished - 2020


  • FinTech, digitalisation, robo advice, MiFID II, portfolio choice, investment decisions, experimental analysis, laboratory


Dive into the research topics of 'Do Investors Rely on Robots? Evidence from an Experimental Study'. Together they form a unique fingerprint.

Cite this