Understanding Reverse Mortgage Acceptance in Spain with Explainable Machine Learning and Importance–Performance Map Analysis

In developed countries such as Spain, where the population is increasingly aging, retirement planning and longevity risk represent major societal challenges. In Spain, in particular, a significant proportion of household wealth is concentrated in real estate, primarily in the form of owner-occupied...

Descripción completa

Detalles Bibliográficos
Autores: Andrés Sánchez, Jorge de, González-Vila Puchades, Laura
Tipo de recurso: artículo
Estado:Versión publicada
Fecha de publicación:2025
País:España
Institución:Universidad de Barcelona
Repositorio:Dipòsit Digital de la UB
OAI Identifier:oai:dnet:ubarcelona__::db8ae8fc192eefc7cc2e7ef9edd7aca6
Acceso en línea:https://hdl.handle.net/2445/229162
Access Level:acceso abierto
Palabra clave:Longevitat
Hipoteques inverses
Longevity
Reverse mortgage loans
Descripción
Sumario:In developed countries such as Spain, where the population is increasingly aging, retirement planning and longevity risk represent major societal challenges. In Spain, in particular, a significant proportion of household wealth is concentrated in real estate, primarily in the form of owner-occupied housing. For this reason, one emerging financial product in the retirement savings space is the reverse mortgage (RM). This study examines the determinants of acceptance of this financial product using survey data collected from Spanish individuals. The intention to take out an RM is explained through performance expectancy (PE), effort expectancy (EE), social influence (SI), bequest motive (BM), financial literacy (FL), and risk (RK). The analysis applies machine learning techniques: decision tree regression is used to visualize variable interactions that lead to acceptance; random forest to improve predictive capability; and Shapley Additive Explanations (SHAP) to estimate the relative importance of predictors. Finally, Importance–Performance Map Analysis (IPMA) is employed to identify the variables that merit greater attention in the acceptance of RMs. SHAP values indicate that PE and SI are the most influential predictors of intention to use RMs, followed by BM and EE with moderate importance, whereas the positive influence of RK and FL is more reduced. The IPMA highlights PE and SI as the most strategic drivers, and RK and BM act as relevant barriers to the widespread adoption of RMs.