Multilingual Topic Modelling of Retirement Preparedness Discourse on Facebook

Abstract

Retirement preparedness is increasingly shaped by socio-economic conditions, institutional communication and public perceptions. Despite the growing use of social media as a platform for discussing financial and retirement-related issues, few research has examined how retirement preparedness is communicated across different socio-economic and linguistic communities. This study aims to investigate public discourse on retirement preparedness and identify key concerns among Malaysia’s B40 and M40 income groups using multilingual social media data. A total of 48329 comments were collected from major Facebook communities, news organizations and media platforms between 2020 and 2024. Comments in English, Malay, Chinese and Tamil were processed using Latent Dirichlet Allocation (LDA) topic modelling to identify dominant themes and discourse patterns. The findings revealed five major themes: (1) government policies and subsidies, (2) savings and withdrawals, (3) cost-of-living concerns, (4) financial planning and investment and (5) family and cultural responsibilities. Differences were observed across language communities among B40 and M40, reflecting distinct cultural perspectives on retirement preparedness. The study demonstrates the potential of multilingual social media analytics to generate actional insights from large-scale unstructured data. The findings can support policymakers and financial institutions in developing targeted retirement communication strategies and evidence-based interventions for diverse communities.