The seventh FinTech Paper is part of a broader survey on the FinTech phenomenon that Consob launched in 2016, in collaboration with several Italian universities, with the aim of exploring opportunities and risks arising from the application of technological innovation to the offer of financial services.

In particular, the study integrates the research strand dedicated to robo advice with an original contribution on the behaviour of a specific group of potential robo advisor clients, i.e. the so-called millennials and post-millenials, who, according to an evidence-based approach, could be taken into account in specific initiatives to protect investors. In particular, the research aims to verify whether an individual's propensity to follow an investment recommendation changes according to whether the advice is given by a human adviser or a robo advisor.

To this end, the document uses data and evidence collected as part of a laboratory experiment, designed and organised in line with the indications of the experimental economy. The results of the experiment, which involved about 180 university students at Luiss, show that the probability that an individual will follow an investment recommendation does not depend on the nature of the advisor (i.e. it does not depend on whether the advisor is physical or digital), but on the gap between the choice made independently before receiving the advice and the choice recommended by the advisor.

In detail, the probability that the investor is willing to follow the advice of the advisor (human or robo) increases if the recommended portfolio coincides with the one previously chosen independently. This evidence could be explained, among other things, by a propensity to the so-called 'confirmation bias' (i.e. the attitude to consider among the available information especially those that confirm pre-existing hypotheses and opinions).

In cases where the autonomous choice differs from the recommendation received, participants seem more inclined to follow the advice of the human advisor and less inclined to follow the advice given by an algorithm. Finally, the results show that the female students participating in the experiment tend to follow the advice received by the physical advisor more frequently if the advisor is a woman than if the recommendation was made by a man.