This article originally appeared in the print edition of Retirement Magazine Vol. 2

Encouraging super fund members who don’t want or can’t afford full financial advice to make the best possible decisions for themselves has emerged as the holy grail of super funds developing effective retirement income strategies.

When it comes to retirement matters, individuals left to their own devices will reliably do nothing. If they do take action, it often won’t be the optimal course for them to follow. Members are only human, after all.

An understanding of behavioural economics concepts and what motivates individuals to act is central to the idea of nudging, and to helping individuals make the best decision, without explicitly telling them what to do.

Shlomo Benartzi, a global pioneer in the field of behavioural economics and co-founder – with Nobel prize winner Richard Thaler – of the Save More Tomorrow (SMarT) program, says that doing anything right now to prompt an individual to move in the right direction is preferable to waiting for a perfect answer or suggestion which may take years to develop.

“If I were to give you what might be the number one lesson from behavioural economics, it is: make the smart choice the easy choice,” Benartzi says. “And another line that would probably be less for the individuals and more for the providers and the advisers: don’t let the perfect be the enemy of the very good.”

In other words, keep an eye on the big picture, and don’t get too hung up on the details. For example, Benartzi says, it’s near impossible to tell if an individual member of a fund should save 14.1 per cent or 14.2 per cent, “but many people should save more than 12 per cent”.

“The fact that we don’t have an exact number to the decimals doesn’t mean that we can’t do much better, and the same would apply to decumulation and producing retirement income,” he says.

“We could keep debating here forever how much you should annuitise and how much should be what, in the US, we call systematic withdrawals. But knowing whether I need to have 33 per cent annuitised or 34 per cent annuitised, if anyone that tells you they know exactly that, they’re just bluffing.”

They’re selling what behavioural economists like Benartzi call “the illusion of precision”.

“We need to get 90 per cent of the way there, with 10 per cent of the effort,” Benartzi says.

“We’re never going to get this to be 100 per cent perfect and it doesn’t really matter, at the end of the day.”

Benartzi says individuals shy away from making decisions they perceive as difficult or complex because they are “afraid of making the wrong decisions, afraid of looking dumb, afraid of admitting that they need help, and they’re a bit lazy”.

“Doing the right things might take a bit of effort.
Doing nothing is the easiest thing, and we often make it easy to do nothing. That’s a big problem.”
– Shlomo Benartzi

“Doing the right things might take a bit of effort,” he says. “Doing nothing is the easiest thing, and we often make it easy to do nothing. That’s a big problem.”

Sitting on a goldmine

Todd Stevenson, Aware Super head of brand, marketing and digital, says data and the application of AI are two essential tools to help combat member inertia, improve engagement and develop effective nudges.

“Super funds are sitting on data gold mines,” Stevenson says. “There’s so much data that we already have, we don’t need more data, we just need to use it smarter.

“We believe in the power of AI, but obviously in responsible application. It is a cultural and capability build that you need, so we are investing heavily into change management [and] building the capability internally to then use AI in responsible ways to drive better outcomes for members: how can we be doing things more efficiently; how can we deliver better outcomes to members through better insights or information that is relevant to those individuals?”

Benartzi says the application of AI cannot be overstated in shortening development and testing times.

“In the old days the way a behavioural economist would try to help people save, we would run experiments. We will try different things. We will try to automatically enroll people – we found that it’s very effective. Australia decided maybe we just make it mandatory – that’s also a possible path, I’m not saying it’s good or bad – but we would run experiments.

“It often took us years to run an experiment. If you think about automatic escalation of saving rates and the Save More Tomorrow program I developed with Richard Thaler, it took us three years to find a plan sponsor to try, three more years to see how it works – you’re talking years. Now there’s a new world out there that is about using an AI to run tests in matters of days or hours versus months or years.”

Benartzi says the field of generative populations is having a massive impact, allowing organisations such as super funds to run trials on member cohorts by creating their “digital twins”.

“We can ask them a lot of things. The accuracy is actually mind-boggling. Now think about it: we can ask a lot of things, nobody would know it. The range of things you might feel comfortable asking, so you could learn, is much broader. You don’t have to go to compliance and legal. There are no real dollars [at risk].

“It’s pretty easy to do it and the speed and cost is something totally new.”

In fact, Benartzi says, running sandbox experiments like this potentially makes the job of getting the real-world green-light easier.

“You go into compliance with the evidence and say, if we did this, we should expect these kinds of results, and only then you’re spending the technology dollars to build the pipes or whatever you need to make it a reality.”

Stevenson says this is a development Aware Super is keen to explore and he says he is “a huge believer in the power of this technology”.

“I’ve seen a lot of the research, and the statistical accuracy that comes with it is exceptional,” he says.

“It really presents us an opportunity to test things at speed, get things to market faster for members, really understand the nuances between members and cohorts, et cetera, and ultimately, it’ll allow us to deliver, in effect, personalised propositions in the future.

“Think about Netflix in terms of how your front screen is very different to mine, is very different to Shlomo’s. That, in the future, will be the experience the super fund will deliver.”

This time it’s personal

Benartzi says it’s difficult for members to differentiate one super fund from another. They all largely do the same things in the same way. And, because of the herding mentality fostered by the Your Future Your Super performance test, they all look the same performance wise, with just a few basis points of return separating the field.

Benartzi says the funds that will separate themselves from the pack in coming years will be those that are “smart about AI; AI is going to be a game changer for the members”.

“They could really move the needle on financial wellbeing,” he says.

“They are obviously going to gain a much bigger market share. Their economics would work better. And those who would kind of be slow and less innovative and fall into kind of the big organisation issues are just going to be left behind.”

Benartzi says studies show that AI projects tend to fail not because the technology is lacking, or that it is applied to addressing the wrong issues, but because organisational cultures do not adapt fast enough to its use.

“It’s not about hallucinations, it’s not about AI not delivering quality output, it’s about the organisations not being able to integrate and adapt to the AI,” he says.

Stevenson says AI will allow funds to deliver, in effect, personalised propositions in the future.”

“We’ll be able to understand individuals and be able to deliver tailored communications, [or] personalised experiences relevant to them.

“You can test interfaces. You might have historically tested an A-B site and run it for three or four months to get the outcome. You can do that now almost daily, you can do that now instantly.

“But exploring those types of scenarios… is much deeper than just UX or UI. You get product development, you can test advice propositions, you can test language before you go and spend significantly on a large advertising campaign.”

Nudge, nudge

Even if technology allows funds to gain deeper insights into members’ wants and needs, there’s still the task of creating effective communications and creating nudges to help members to get where they want to go. Benartzi says that to create a “good” nudge, “you need to know what’s a good outcome”.

“Sometimes, we can have an objective metric of people saving enough. We could run some income replacement calculation and say, ‘Are they on the path to successful retirement?’,” he says.

“I do think good nudges should not override the important role of freedom of choice and personalisation. And what I mean by that is, maybe the nudge makes it very easy to go left, but it should be easy for people who don’t want to go left to say, hey, I want to go right.

“If you think about countries like the UK, where people are automatically enrolled in a DC plan, we’ve made it very easy to start saving, that’s great; but we should also make it very easy to opt out, for the single parent with three kids and five maxed-out credit cards who cannot save now. That’s about making it easy to choose; and even if there is a good default, it might not work for everyone.”

Benartzi says the development of nudges must recognise the world is a lot more complex than it was just a couple of decades ago, and that the financial decisions individuals must make as they move into retirement are correspondingly both more complex and more idiosyncratic.

“A good nudge is a nudge that recognises individual differences,” he says.

“That’s something that we didn’t pay enough attention to as behavioural economists 30 years ago, because the technology and the scale to personalise was not there. Saying everyone should save two per cent more every year was an incredibly good starting point, but nowadays we have a lot more data, and not everyone should save at the same rate.”

Stevenson says a challenge in developing effective and personalised nudges for members at any point through their life, but especially in relation to retirement, is that “retirement is so far away for people, they can’t visualise what ‘good’ looks like for them”.

“They don’t know what they need in retirement,” he says.

“If we can help individuals identify what their goal is for their future – and that can be done, again, using behavioural economics and envisioning of the future – you can work backwards from there.”
– Todd Stevenson

“But if we can help individuals identify what their goal is for their future – and that can be done, again, using behavioural economics and envisioning of the future – you can work backwards from there.”

Stevenson says the goal for a super fund isn’t one generic nudge for an entire cohort of members, but rather “personalised nudging at scale, driven by AI and data”.

“It’s proactive rather than reactive and to help people make good decisions, or even circumvent bad decisions… framing is really important,” he says.

“An example is that members will respond much better to stories, outcomes, emotions and examples – an extra $100,000 in retirement will land much better than a point two per cent fee reduction. And then that dynamic framing – personalising those messages depending on the individual – would be really important.”