“Imagine a parade of 10,000 people marching past your house every day, each holding a retirement cake and a balloon. That's a lot of calories & balloons and that is only in the United States!”¹

Projecting forward, today’s 120M+ active participants in retirement plans in North America will be retiring over the next few decades. More than 70% of those will not have guaranteed income for life from private pensions. It is no surprise that decumulation - the problem of drawing down one’s savings in retirement - has recently become a topic of interest to researchers and economists. The enormity of the problem highlights the importance of this issue to both public policy experts as well as the financial services industry. 

There are many reasons why decumulation is deemed to be the hardest problem in finance, but at its core it is all about creating certainty amid uncertainties: 

• Converting assets to guaranteed income while managing the unknown investment and longevity risk
• Few one-time decisions with long-term consequences and very little opportunity to get feedback and reverse those decisions

1 - Image by rawpixel.com on Freepik

A wrong decision can lead to reduced retirement income at best and catastrophic outcomes at worst. It is imperative that individuals have a somewhat clear understanding of their situation before making such decisions. Hence providing financial literacy on decumulation - one based on behavioral science - is crucial to the financial well-being of future retirees. 

While applying psychology to the accumulation of assets has been successfully yielding higher participation and savings rates, the complementary problem of drawing down one’s savings in retirement has been relatively unexplored until recently. Based on fairly recent research² on psychology and behavioral finance, human cognitive biases pertaining to decumulation can be broadly categorized into loss aversion, temporal discounting, self-control, fairness or psychological ownership. “Cognitive biases are systematic errors in thinking that can lead individuals to make irrational or suboptimal choices and are particularly prevalent in situations where there is a higher level of complexity and uncertainty, such as decumulation.”¹

Financial literacy built on psychological insights can be a powerful tool in the toolbox of regulators and retirement plan administrators to improve the well-being of tens of millions of retirees at very little cost. As an example, let’s look at one specific challenge in the decumulation space commonly referred to as the annuity puzzle. “It refers to the phenomenon where people tend to avoid buying annuities, which represent a rational product choice for ensuring long-term financial security in retirement. This outcome can be attributed to several factors, including a lack of understanding of annuity products, a preference for flexibility and control over one’s finances, and a tendency to focus on short-term gains rather than long-term planning.”¹

The annuity puzzle has been studied extensively by academia. We briefly highlight two pieces of research that we find particularly interesting and relevant to our topic. In a comprehensive study by Brown et al³, authors posit that the annuity puzzle is partially attributed to framing biases. They have discovered that people do favor annuities if they are presented in a “consumption frame” that considers the consequences for lifelong consumption, rather than "investment frame" that focuses on risk and return. 

In another study, Schreiber and Weber⁴ describe the time inconsistency behavior as it comes to annuities. In a nutshell, people favor annuities when they are younger but dislike it as they get closer to retirement age. Such behavior is partially attributed to hyperbolic discounting, ownership and fairness biases. 

Research has shown that we can test for many such biases and correct for them as we engage with members early in their careers. Financial literacy can play a crucial role in overcoming the biases which impede an individual’s ability to make rational choices and decisions. In order to be truly effective, a multi-pillar framework for digital financial literacy – one that is built on member engagement - needs to be reinstated which consists of the following pillars:

1. Build an effective financial literacy program using human-friendly tools and relatable communication language

2. Engage with members through their entire career to learn about their biases, worries, needs and expectations

3. Implement just-in-time learning that enables users to apply accumulated knowledge right at the point of decision making, which tends to be a point in time when individuals mostly rely on “human flaws”

The three components are different from one another and one does not necessarily lead to the other but at every stage there are small wins to be made which - over time - translates into big leaps. 

This can be easier implemented for members who stay with a single retirement administrator or solution provider by keeping the engagement alive over many years or decades. The administrator’s engagement strategy incorporates these steps into their financial literacy program to aid members with gentle nudges along the life income journey. 

This all sounds great in theory but how do we apply it in practice? Thankfully, today’s technology has given us what could only be called magic just a couple of years ago. The new engagement models, rendered possible by technology, allow for a two-way conversation across a range of topics in finance. The engagement gives us an opportunity to learn about financial knowledge & perception of participants at a low cost and at scale. It helps with personalization of the message and member education in the right area, at the right time and with the right language. One could almost say literacy becomes a circular process! 

Through engagement strategies, users learn about finance and we learn about their needs, desires and expectations which again feeds back to products, education and member engagement provided. And AI will be at the front, back and center of it with some of the following implementation steps:

• Conversational AI to test for certain behavioral biases at scale
• Machine learning (ML) for segmentation of members into different categories
• Personalization of further engagement steps based on the feedback received 
• ML-based optimization to offer appropriate products and help with decision making

2 – Image by DALL-E

Given that research in the field of psychology and behavioral sciences has resulted in significant improvements for individuals in certain areas of finance, transferring those insights to the decumulation space is certainly within reach. However, to make this happen, a concerted multidisciplinary approach is required based on a more holistic financial literacy framework which is brought to life with the latest tested and secure technology. 

We have used ChatGPT - with citations - for this article mostly to get it interested in the decumulation topic. Out of curiosity, we also asked ChatGPT an important question about the future of this field.

Do you believe the challenges in decumulation can be better addressed by AI or humans?

The challenges in decumulation are best addressed by a combination of AI and human expertise. While AI can provide valuable insights and recommendations based on data analysis, human judgment and decision-making are necessary for navigating complex and emotionally charged retirement income planning decisions. A combined approach that integrates AI tools and human expertise is likely to be the most effective way to address the challenges in decumulation.