Nudging kiwisaver

New Zealand’s kiwisaver scheme has proven to be a political success. Created in 2007 in response to an impending slump the number of people in work relative to people in retirement, funds have grown by leaps and bounds. The plan has now become a politically and fiscally indispensable part of New Zealand’s future collective retirement savings. As of March 2013, membership had reached 2 million, or just under half of New Zealand’s total population of 4.4 million [1]. Since 2008, members have saved over $14 billion, or 40% of the $35 billion in New Zealand super funds [2].

There’s a number of changes one could make to kiwisaver to make it a better fit for the way people relate to their money and retirement savings in practice. Much of basic economics assumes people are ‘homo economicus’, an entirely rational human capable of perfect market action, but knowledge about human psychology – our biases, systematic mistakes, and evolutionary ‘flaws’ -  can help us design a savings system better suited for homo sapiens. There are a number of features in kiwisaver that approach an intuitive, homo sapiens model, but several improvements could be made. First, there may be more cost-effective ways to incentivise people to join the plan - and commit an adequate level of savings to it - than putting $1000 towards their savings at the start. Second, gradual increases in contribution rates, especially timed with increases in income, could help people to reach a higher, more realistic retirement savings level. Third, there does exist an argument for a publicly-run kiwisaver scheme in competition with the current private ones, although the argument rests more on economics than cognitive psychology.

The graph shows the total payments made for each financial year since the year ending 2008. Payments have continued to climb with membership since the scheme started.

In their book Nudge: Improving Decisions About Health, Wealth, and Happiness, Thaler and Sunstein argue that human cognitive biases prevent us from behaving in ways conducive to our own best interests, and that often – particularly in an economic system based around homo economicus, or purely rational beings, operating in a free market – we have outcomes that are less than optimal for both individuals and for society. It is well recognized in cognitive psychology that we have a tendency toward making cognitive errors in particular defined ways. Thaler and Sunstein’s argument is that where possible, economies and societies should be redesigned in such a way that – without preventing people from making their own choices – it becomes easier for people to make good choices and easier to avoid making bad ones.
One simple example of this kind of design can be found in retirement schemes like New Zealand’s kiwisaver. The kiwisaver scheme gives employees an immediate 250% return on their investment and continued return until they retire, leading many personal finance advisers to describe joining the scheme as a “no-brainer”, despite the fact that law prevents savers from taking their money out of their kiwisaver account until they reach retirement age. If people starting a new job are, by default, outside the scheme and must opt-in to take advantage of it, they have the freedom to decide whether to invest in the scheme or not, but doing nothing will leave them out of it. On the other hand, if people starting a new job are, by default, inside the scheme, but can opt-out if they wish, people still have freedom to decide whether to belong to the scheme, but doing nothing will allow them to take full advantage of the scheme. In each case, individuals must implicitly or otherwise make a choice about belonging to the scheme; in each case, individuals have the freedom to decide whether they’d like to invest in the scheme or not; but the latter case will lead to much more money saved in kiwisaver and likely much more funds available to people when they retire.

Thaler and Sunstein’s helpful (if slightly contrived) mnemonic NUDGES describes what they believe “choice architects” can do to design better choices for people:

•  iNcentives
• Understand mappings
• Defaults
• Give feedback
• Expect error
• Structure complex choices

The current structure of kiwisaver results from the usual mix of opposing ideals and political decision-making from New Zealand’s Labour government from its introduction to 2008 and then the National Party government from 2008 until the present. But the question can be asked whether the system measures up to the standards described by Thaler and Sunstein’s “NUDGE” mnemonic.

Immediate financial incentives have in some form been present in the scheme since its inception. Although kiwisaver savers potentially accrue a return at market rates throughout the life of their investment, a special bonus is obtained immediately at the outset. First, government offered an initial “kick-start” payment of $1000 when individuals join the scheme. Second, government and employer subsidies are added at the same time individuals contribute to their scheme. This last feature has been tweaked a number of times since 2007; the National government reduced required employer subsidies from a maximum of 4% to 2% in 2009, but more recently raised those subsidies to 3%, while lowering the government subsidy. The extent to which financial incentives such as these operate as real psychological incentives is questionable, and there may have been better ways to incentive participation in the scheme using more immediate awards. Thaler and Sunstein explicitly used Kiwisaver to point out the questionable power of purely economic incentives (relative to intelligent defaults). Despite the $1000 kickstart offered by the government for kiwisaver, after merely six months, the automatic enrolment for people starting new jobs was the main method of enrolment, despite the kickstart enticing people to join through active enrolment [3]. If more immediate incentives are more powerful, perhaps a “rewards” scheme, as is popular with credit card accounts, may be more cost-effective. Although less government money goes into the scheme, more immediate incentives could be more powerful and ultimately lead to more money in the scheme altogether. However, such “bribes” have their own problems; they could undermine the extent to which the scheme is taken seriously. The kickstart also functions as a direct investment in each individual’s retirement fund.

People should be able to understand mappings of choice labels and descriptions into outcomes. Too often, choice presentations and labels do not easily and intuitively suggest outcome. Understanding mappings in choices with kiwisaver could be a significant challenge for potential savers. Individuals must choose between an array of index-tracked and actively managed funds without necessarily understanding how funds are invested. In fund promotional brochures, past fund performance is always indicated next to a disclaimer Past performance is no indication of future returns, leading consumers to either rely on them anyway or wonder about the point of such indications. Scheme providers such as the ASB Bank describe their own schemes using labels describing points in a continuum from “Conservative” to “Growth”. Still, only a few schemes (the Gareth Morgan Scheme is one exception) allow investors to see exactly which stocks and bonds their money is invested in, which might constrain the ability of more financially-literate investors to make their own best possible choices about funds in which to invest.

The power of an intelligent default opt-in for enrolling members in kiwisaver has been described above. A default opt-in could have opted in all employees, with an option to opt-out if they wished, and would have seen even more powerful results. This kind of automatic opt-in could still be applied at present, though it does seem that the majority of working-age New Zealanders are now in the scheme. However, Thaler and Sunstein extensively discussed examples of setting intelligent defaults for investors’ contribution level and the scheme into which they’re enrolled. Thaler and Benartzi’s successful “Save More Tomorrow” private savings scheme enrolls members at an initial low rate – say, 1% - and increases members’ rate in sync with their pay rises. In that way, members commit to very little in the short term, making the scheme easier to start, but in the long term, build a worthwhile savings rate and never have to adjust to a lower income as a result. In a company where the scheme was tried, just 25% of employees increased their contributions rate by 5% when recommended to by an advisor, but 78% of those who declined agreed to the “Save More Tomorrow” gradual rate increase scheme. Three years later, while the first group were contributing an average of 9% to their savings scheme, the second group – initially contributing just 1% - were contributing 13.6%, very close to the 15% recommended as optimum for the whole group by advisors. Apparently the plan has caught on, because by 2007, 39% of large employers in the USA were reported to use some variant of automatic rate increase.

Further default setting concerns for retirement funds include appropriate exposure to risk (high-growth, high-risk investments for people far from retirement, to low-growth, low-risk investments for those about to retire) and choosing a particular default fund. Thaler and Sunstein explained that Sweden’s privatized social security model, introduced last decade, allocated members to a carefully chosen, government run scheme, run in the model of Sweden’s previously successful government funds, if they failed to pick one for themselves. It turned out this scheme performed better, than the average of schemes chosen by savers themselves. In the New Zealand system, when a new member joins kiwisaver, if employers do not choose, rather than allocating the member to a carefully chosen scheme, government randomly allocates members [4] to one of a group of several private schemes chosen by the government. Such random allocation, while relying on private schemes, blunts market competition for quality, an essential part of any privately-run investment offering. Care should be taken when encouraging members to choose their own scheme, as members don’t necessarily choose better than a well-designed default; any default should avoid the kind of market inertia which could arise from randomly allocating members to private schemes. Morningstar’s most recent Kiwisaver Performance Survey report [5] named 5 best-performing kiwisaver providers over the previous 3 years or 5; none of these schemes were government defaults; though the report doesn’t mention any impact which management fees might have on investor returns. One might wonder whether private, government-appointed default scheme providers have less incentive to provide well-performing schemes.

Feedback for kiwisaver schemes might include clear and transparent information about returns and management fees. Savers ought to be able to easily compare and choose competing schemes. As this article is primarily focused on human factors related to kiwisaver, it’s beyond the scope of this article to examine deeply the way that different schemes can vary. Successive governments have tweaked kiwisaver disclosure rules.

Kiwisaver schemes are designed with some kind of expectation of error, at least in terms of potential mistakes made by clients and managers. Rules specify who is qualified to give different kinds of kiwisaver advice – as I write, the New Zealand Herald reports those rules have just been relaxed somewhat, though apparently only temporarily [6].

Structuring complex choices is the other topic Thaler and Sunstein discussed extensively with regard to government savings schemes. Investment experts know a number of facts not well understood by many kiwisaver savers, from basic principles like “diversifying”, to appropriate exposure to risk, to expert opinion in judging performance of investments. Currently the sheer number of schemes offered across all providers makes for very complex decisions on the part of savers. Savers need to be given an easy, clear, process to choose the scheme that is best for them, taking into account issues such as risk profile.

Inevitably, design of government-run savings schemes will be influenced by political realities. There are also practical constraints, such as the level of government and employer contributions deemed wise. Other legitimate debates exist about the merits of individualized savings schemes such as kiwisaver compared to schemes provided for by general taxation and paid out at a specified rate, such as New Zealand’s current Superannuation system. But within the constraints of an individual contribution system, where the amount paid out is determined by the amount invested and market returns, there remain a number of ways in which the system can be improved. These include tweaks to government incentives, a graduated, automatic, and optional rise in contribution rate with rises in income, and close scrutiny of default plans with consideration of a publicly-run option.

---

[1] Dwyer, M. (2009) The Place of Kiwisaver in New Zealand’s Retirement Income Framework. http://www.cflri.org.nz/sites/default/files/docs/RI-Review-2013-KiwiSaver.pdf, p6. Retrieved 18 August 2013.

[2] Douglas, C. (2013) KiwiSaver Performance Survey June Quarter 2013. http://www.morningstar.co.nz/s/documents/KiwiSaver-Survey-30-06-2013.pdf. Retrieved 18 August 2013.

[3] KiwiSaver evaluation reports: KiwiSaver Six-monthly report 1. http://www.ird.govt.nz/aboutir/reports/research/report-ks/research-ks-2007-12-31.html. Retrieved 18 August 2013.

[4] KiwiSaver: Choosing your KiwiSaver scheme. http://www.kiwisaver.govt.nz/new/providers/

[5] http://www.morningstar.co.nz/s/documents/KiwiSaver-Survey-30-06-2013.pdf. Retrieved on 18 August 2013.

[6] David Chaplin, The New Zealand Herald, 15 August 2013. “Code on the road - new rules for advisers” http://www.nzherald.co.nz/david-chaplin/news/article.cfm?a_id=616&objectid=10912346. Retrieved on 16 August 2013.

Public Perception

Over the next few years I’ll be completing a PhD in social psychology at the University of Southern California. I plan to focus on projects using computational modeling techniques (like neural networking and coherence-based modeling) to model systems in social psychology. These could be systems at the individual level – social perception and attitudes – or involve some form of group modeling as well. I hope to focus on moral psychology – the science examining how moral judgments are made – as well as how moral psychology relates to the political decisions, behavior, and allegiances that people tend to form in contemporary society.

This blog takes the approach that understanding our individual psychology properly is essential for understanding how we can live together well. Many people take this approach in their personal lives. In this blog, I am to show what we learn from science about our psychology tells us about how we ‘live together well’ in a public sense – in our economy, in our democracy, and our society. That’s how this blog gets its name – here you will read about how we can take what we know about how we perceive ourselves and each other as individuals, and apply it to help design a better public sphere.

Topics will undoubtedly draw inspiration from my research, as well as the work of other social psychology researchers, but will have a broader focus in psychology while aiming to apply psychological principles in economics, society, public policy and politics; particularly in current issues. I’ll mainly focus on the United States and New Zealand contexts, since I am most familiar with those.

The first post, up in a couple of days, is about suggestions authors Richard Thaler and Cass Sunstein made in their 2008 book Nudge, based on behavioral science and economics, to improve employee savings schemes like kiwisaver.