The CFA Institute Research and Policy Center has released new research exploring how AI and machine-learning technology can address critical issues facing the global pensions industry.
The association writes that demographic shifts, underfunding of defined benefit pension plans, inflation, rising inequalities, and gaps in financial literacy, are creating an increasingly complex pensions environment and more risk of pension inadequacy for retirees.
The research, “Pensions in the Age of Artificial Intelligence,” explores whether technology can offer solutions to address key issues across various parts of the value chain in pension management. It includes case studies and expert interviews from markets including Australia, Japan, the United Kingdom and the United States, and surfaces examples of where the integration of new technology can empower pension trustees as well as beneficiaries to make better informed decisions.
Olivier Fines, CFA, Head of Advocacy for EMEA at CFA Institute, comments: “AI can add more than just operational efficiencies. From onboarding new members to account management, plan governance, investment management, and decumulation strategies, AI and machine learning can play a positive role in addressing key issues facing the pensions industry. However, transparency and robust governance in the use of the technology will be vital to build trust and rapport with pension plan members.”
“As investment firms and plan sponsors increasingly integrate the technology in their processes, collaboration with technology providers and regulators will be important to ensure that workflows consider the specifics of individual pension plans, and respect their operating models, fiduciary duty, and the regulatory framework they operate in. This is why we believe AI should support, not replace, human decision-making and why it will be critical to set clear objectives and benchmarks for evaluating model effectiveness.”
“AI and machine-learning technologies may allow providers to offer more personalised services and dynamic support for plan members. We believe AI should play a major role in balancing personalisation and simplicity, accounting for individual needs and varying levels of financial literacy. This should foster increased member engagement and promote financial wellness. AI can also assist trustees and advisors in navigating the crucial decumulation phase for plan members.”
Key Findings:
· Enhancing Personalisation, Efficiency, and Accuracy: AI applications are diverse and consideration must be given to how to best use AI to enhance overall retirement security for pension plan members. Enhancements will require targeting specific areas of the pension ecosystem that will contribute the most value for the unique needs of each pension fund.
Member Engagement and Financial Literacy: Implementing AI for member onboarding, communications, reporting, and retirement planning could enhance overall member engagement, boost financial literacy, and support pension plan members throughout their retirement life cycle.
Pension Plan Governance: AI technologies can enhance pension plan governance by facilitating multi-stakeholder interactions, reducing administrative tasks, and aiding pension boards with decision-making. This includes improving optimisation of investment strategies and prompt resolution of member issues.
Investment Management: AI and machine learning models can boost the analytical capacities of portfolio managers, enhance actuarial analyses of pension fund risks, and keep market trend assessments up to date. These technologies may be especially useful for analysing private markets and data related to sustainable investments.
Predictive Analytics and Actuarial Assumptions: Advanced machine-learning techniques may enhance actuarial assumptions and predictive analytics, improving asset/liability management and pension derisking strategies. Defined contribution plans may benefit from personalised strategies across the lifecycle of each individual investment plan, with accumulation and decumulation strategies based on member behaviour predictions.
