Who is unhappy for Brexit? A machine-learning, agent-based study on financial instability
Source of Publication
International Review of Financial Analysis
© 2020 In this paper, we assess the happiness cost of Brexit in the UK and the EU, using data from the Gallup World Poll. We implement a two-stage learning machine, using a naive Bayes classifier to extract happiness preferences of the population and then passing these onto an artificial neural network of attributes to generate dynamic happiness functions for each household, on an agent-based modelling framework. We find that there is a significant long-run cost in terms of both happiness and unemployment, which primarily affects the most vulnerable portion of the population. In addition, despite the expected instability in City's financial centre, the UK financial sector seems to be well equipped to deal with the repercussions, thus minimising the welfare costs for the country. Our findings extend the discussion of the economic costs of Brexit, by adding the welfare cost of the ensuing financial instability.
Agent-based finance, Banking crises, Brexit, Happiness economics, Machine learning, Naive Bayes classifier, Neural networks
Polyzos, Stathis; Samitas, Aristeidis; and Katsaiti, Marina Selini, "Who is unhappy for Brexit? A machine-learning, agent-based study on financial instability" (2020). All Works. 3994.
Indexed in Scopus