Think Again
Behavioural Finance & Retirement Pensions
It is paramount that retirement income systems and the advisers, trustees and other fiduciaries responsible for their management strike a fine balance between individual freedom and subtle nudges or paternalistic interventions. While a life cycle savings economic model provides a sound foundation, there is a need to consider the key decision points from a behavioural perspective if outcomes are to be best directed towards promoting an adequate and dignified retirement for ageing populations.
In the early 1950s, Nobel prize winning economist Franco Modigliani and student, Richard Brumberg theorised the Life Cycle hypothesis, suggesting that individuals make spending and saving decisions based on the resources available to them over their lifetime and on their current stage of life. Modigliani asserted that individuals made rational spending and saving decisions to build up assets at the throughout the middle stages of their working lives, before drawing down on these in retirement. It is a model of lifetime self-sufficiency and responsibility with roots in neo-classical economic theory. It is an economic model which is rational, reasonable and removed from reality.
Few would disagree with the notion that it makes good economic sense to save for retirement — yet a cursory observation of the world will confirm that there are an awful lot of people whose decision-making behaviour is inconsistent with their rational views. Scarcity of income for saving, and legitimate expectations of dependence on state funded or occupational age pensions or family support in retirement are important factors in shaping decision making. However, there are also more fundamental behavioural traits and biases which influence decision making when it comes to retirement income planning. It is important that individuals, fiduciaries, and those responsible for the design of retirement policies are aware of these behavioural traits and fashion retirement decision making in a manner which accomodates such traits towards the best interests of individual retirees.
Dividing decision making
The fact that individuals often fail to make decisions to smooth their spending and consumption over their lives in saving for a self-sufficient retirement has long been known. Paternalistic systems have developed to accommodate this, utilising family structures, governments, employers and fiduciaries such as trustees and advisors to make decisions on behalf of individuals. This division of decision making between the individual and a fiduciary is paternalistic, however it may be beneficial in promoting more objective decisions which promote the best interests of retirees by avoiding detrimental behavioural traits in decision making.
This division of decision making responsibilities between individuals, fiduciaries and governments has typically been reflected in the three-pillar model which remains common across much of the world. Governments rely on public policies using taxation powers to finance age pensions, fiduciaries are responsible for the management of occupational pensions and individuals remain responsible for private savings.
While such a characterisation is neat, it is more nuanced and the subject of continued change. The role of individual decision making in the management of occupational pensions is playing an increasingly important role, particularly in Northern America and Australia where individual control over investment and other decisions is on the ascent.
Irrespective of the distribution of decision making powers, it is important that decision making is prudent and responsible in ensuring that retirement incomes are adequate, manage risk appropriately and costs are efficient. Neo-classical economic theory provides rational models such as that proposed by Modigliani to inform what such decision making might look like, yet recent advances in behavioural science have shone a light on why individual decision making is often far removed from the rational agent model in practice.
Behavioural Finance
The human mind is wonderfully complex, and this article has no intention of attempting to reproduce the insights which behavioural scientists have illuminated workings of the human mind. However, it is worth briefly touching on some basic characteristics of the human mind which have profoundly unsettling yet familiar implications for rational agent economic models of decision making behaviour.
Our mind is of limited capacity, and has developed wonderful ways of efficiently processing information to make decisions. The associative, intuitive and heuristic adaptability of our minds enables us to avoid paralysis from information overload that a strictly considered and logical approach would entrench. Yet, the brilliant efficiency of the associative, intuitive and heuristic mind has a downside.
Behavioural biases can creep into decisions which would result in better outcomes if they followed a strictly considered and logical approach. There are a broad range of biases which have been scientifically proven, and we’ll only consider a few which are relevant for retirement income decision making:
The popular publications of Nudge from Richard Thaler and Cass Sunstein in 2008 and Thinking, Fast and Slow in 2011 by Daniel Kahneman (the first psychologist to win the Nobel prize in economics) have helped to popularise Behavioural Economics. Behavioural finance extends from the same behavioural science foundation, connecting traditional models based on neo-classical economic theory with the findings of psychologists in behavioural science.
If understood and applied effectively, behavioural finance can assist in promoting better decision making to ensure that the underlying objectives of the rational agent model such as Modigliani’s are met.
There are also challenges in balancing behavioural finance with the neoclassical model, in ensuring that decision making can accommodate psychological realism without collapsing into a mess of special cases. To achieve this balance in the context of retirement planning it can be useful to characterise the key aspects of the rational agent retirement planning and look at how behavioural biases might interfere and importantly how these biases might be managed.
Retirement planning decision making
Governments, fiduciaries and individuals make decisions about retirement income. The distribution of decision making responsibility differs between nations, yet there are a series of central themes which when combined can be said to represent retirement income decision making.
In making decisions in these key areas, there are a broad range of behavioural themes which might afflict rational decisions. Themes of consistency, over-confidence, anchoring, social herding, loss aversion, framing and emotion are caused by underlying behavioural biases which in turn tend to skew rational decision making by both individuals and agents such as fiduciaries or advisors.
If we were to consider the expected decision making of a rational individual in these key areas, we could expect the decisions to correlate somewhat to those in the chart below:
However if we were to look at how the decisions might be formed in the case where an individual exhibited some behavioural biases, then it might look something like this:
The remainder of this article is directed towards understanding how these biases might impact of the key retirement income decisions made by individuals and fiduciaries, with the hope of promoting the development of approaches to improve retirement income system design and decision making by individuals and fiduciaries.
Engagement & control
The level of control that individuals have over their retirement savings is a decision which is becoming increasingly important. It is essentially a decision (or non-decision) about how much of the decision making responsibility the individual wishes to assume. This is particularly prominent in the occupational pension and private savings contexts, with the availability of products and services with various levels of control and flexibility.
Decisions concerning the level of control an individual has over their retirement income generally relate to the balance between an individual and a fiduciary such as a trustee or an adviser. There is no correct level of control, rather decision makers should make decisions based on aligning the level of control with the level of engagement of the decision maker.
In making decisions about the level of control which the decision maker intends to exert, there lies a challenge in ensuring that behavioural biases don’t afflict the decision maker to choose a level of control which is not aligned with the level of engagement. Research in behavioural science has identified traits which are common in individuals and may contribute to decisions which misalign the level of control over a fund or product and the level of engagement of the decision maker.
The Dunning-Kruger Effect is the tendency for unskilled individuals to overestimate their own ability, while experts underestimate their own ability. When making decisions about the level of control over retirement income planning, there may then be a bias towards individuals choosing products which provide a large amount of flexibility and control to the individual, when the individual does not have the required skills to effectively make prudent decisions about their retirement income planning. This may also explain the allure of product which provide a high degree of control and flexibility. While a high level of control may be appropriate for many individuals, there is certain to be large segments of the population who do not have the interest or aptitude to involve themselves in the complexities often involved in diligent retirement income planning.
The IKEA effect is the tendency for individuals to place a disproportionately high value of objects that they have contributed to the construction of, regardless of the quality of the outcome or result. In the context of decisions over the level of control that an individual has over their retirement income decisions, there is a risk that such tendencies may see individuals who have chosen to exercise a greater level of control over their retirement planning being unable to objectively assess the performance and hence value relative to alternate options. This may result in individuals remaining in self-managed accounts or similar occupational pension products longer than they objectively and rationally should.
This overconfidence and disproportionate valuation of adopting a high level of control over managing retirement income savings is often irrational yet is increasingly common. Where retirement income products which require a high level of engagement and control are issued, it may be prudent to establish ongoing obligations on fiduciaries to ensure that there is not an unacceptable risk being posed to the individual’s retirement savings because of the disengagement or negligence which afflicts many in a busy world.
In practice, this might look at monitoring the frequency of trading, investment decisions or even viewing information. If there is a significant change, it may be prudent for the fiduciary to engage with the decision maker to gently remind the decision maker that they haven’t been engaging enough. Similarly, frequent viewing, or switching between high level managed products might indicate that the decision maker is incurring unnecessary costs and might be more suited to products which require a higher level of tactical engagement and control.
Rate of savings
The rate of savings for an individual is a central aspect of retirement income decision making. Indeed, it is at the centre of Modigliani’s theory of logical, self-interested decision making. While there are mandatory minimum savings rates for occupational pensions imposed in some countries (like Australia), individuals typically have the ability to decide on a savings rate or accept a default rate. Decisions about the level of savings are fundamental to retirement income planning, and similarly remain susceptible to behavioural biases which may result in irrational and undesirable outcomes.
The Peltzman Effect is the tendency for individuals to take greater risks when perceived safety increases. Where individuals perceive a degree of safety in retirement incomes due to expectations of government age pensions or family dependency, they may be more likely to make riskier decision such as not saving adequately or at all where the perceived safety of a state funded pension is assumed.
Mandated or default rates of savings have many positive aspects, but may lead to perceptions of safety in that the rate of savings is appropriate, even where it not align with individual needs or expectations for retirement. The risk of inadequate savings due to a perception of safety can however be addresses relatively easily.
Effective projection and communication of what reliance on a state funded pension or a minimum default rate of savings may look like in terms of income and resulting impacts on lifestyle may be quite effective in calling out misconceived safety. Focusing on projections of retirement incomes and the resulting lifestyle changes that might be required are likely to be an effective way to motivate individuals to save at rates more consistent with their desired lifestyle in retirement.
Preservation
The preservation of saved assets is the alternate side of the savings equation. Decisions about savings and spending are intertwined so tightly that they can almost be considered two sides of the same coin. Spending behaviour and decisions directly impact the level savings, and vice versa.
Preservation is typically imposed on retirement savings until retirement, with individuals unable to access savings until retirement, however there are exceptions to this. The preservation of savings takes on additional significance in the retirement phase, as a key factor contributing to longevity risk. Behavioural biases exist which may affect the ability of decision makers to effectively preserve retirement savings in the self-interest or best interests of retirees.
Hyperbolic discounting is the tendency for individuals to preference immediate pay-offs rather than later payoffs. A dollar today is valued higher than five times the amount next month. Essentially, choices and decisions are inconsistent over time, with individuals making decisions today that their future self would not despite using the same reasoning. In many ways, hyperbolic discounting can be regarded as the central behavioural bias which retirement income policies and systems intend to manage. The fact that individuals place a greater value on the immediate spending over future needs.
Hyperbolic discounting has also been shown to be particularly pronounced in circumstances of scarcity. When an individual is exposed to circumstances where time or money is perceived to be in short supply, there is a tendency to focus decision making bandwidth towards the immediate needs such as spending today rather than saving for retirement. This exposes individuals in circumstances of financial difficulty to an increasingly difficult position, and at greater risk of ignoring future strategic decision making (such as preserving savings) to focus on perceived short term or immediate needs.
Retirement income systems are generally designed to ensure that savings are preserved, and do a reasonable job of managing the tendencies we have towards hyperbolic discounting. Yet, there remains an important role for public policy makers and fiduciaries to ensure that preservation decisions are adequate to manage longevity risk. This could well involve prioritising a base level of annuity type products in retirement and discouraging lump sum payments and commutations which increase the risk of hyperbolic discounting seeing accumulated savings being exhausted prematurely.
Asset allocation
The investment of retirement savings is absolutely central to the role which fiduciaries and (increasingly) individuals play in planning for retirement. The rational agent model of self-interest dictates that investment decisions should be directed towards maximising the long term financial performance of investments, in a manner appropriate with the horizon and risk profile of the individual beneficiary.
Investment decision making consists of strategic and tactical aspects, and is the domain of significant complexity and maturity within investment management entities and fiduciaries. While investment managers are not immune to behavioural biases, it is critical that particular attention is paid to ensuring that the retirement income products which allow individuals significant control over the strategic and tactical asset allocation are not compromised by the inherent behavioural biases which can impact investment decision making. While there are a large number of such biases which make for a messy view of investment decision making, there are couple which are particularly interesting and relevant.
The disposition effect is the tendency to sell an asset that has accumulated in value and resist selling an asset that has declined in value. Where individuals are permitted or encourages to control the tactical investment decision making, decisions can be made to buy and sell specific assets such as equities, bonds or units in a trust. This enhances the possibility for investment decisions to be affected by the disposition effect. The greater transparency as of buy and sell prices for particular assets increase the likelihood that individual decision makers will irrationally hold on to poor performing assets rather than cutting losses and disposing of performing assets prematurely.
Public policy makers and fiduciaries need to be particularly mindful of these risks when designing regulations and products which permit an individual approach to managing investments. Fiduciaries are well placed to monitor individual investment decision making, and may be able to identify such patterns in trading activity, and communicate this to individuals controlling strategic or tactical investment decisions.
The gambler’s fallacy also presents a behavioural challenge to investment decision making. It is the tendency for an individual to think that future probabilities are altered by past events, when in reality they are unchanged. The fallacy arises from an erroneous conceptualization of the law of large numbers. For example, “I’ve flipped heads with this coin five times consecutively, so the chance of tails coming out on the sixth flip is much greater than heads.”
The same principle is often observed in the decisions of individuals, with a tendency for investment decisions to be directed towards investments which have performed best in the recent past. To agree, this bias can be a self-fulfilling prophecy, with the increased demand for well performing investments impacting on the value and therefore performance. However longer term it’s important that a broader range of economic fundamentals are considered when making investment decisions. Fiduciaries may be better placed to do this than individuals, and thought needs to be given to extending the regulation of investment decision making beyond ineffective warnings in disclosure rules.
Zero-risk bias is a preference for reducing a small risk to zero over a greater reduction in a larger risk. In the context of asset allocation, such a bias can result in a tendency for individuals to invest in assets perceived as being risk free (such as cash) instead of alternatives which provide a greater return for a similar level of risk. This is particularly relevant for retirees who typically have a risk profile with a lower appetite for investment related risks. There is the real possibility that a low risk appetite is irrationally afflicted by a tendency to prefer options perceived as being risk free. There are measures which can be effective in identifying where cash holdings may be over invested in due to a perception of cash being a zero-risk investment.
The Mere exposure effect is the tendency to express undue liking for things merely because of familiarity with them. Asset allocation may involve the selection of equities or other investments which are overvalued due to a familiarity of the decision maker with the particular investment. The financial media is the source of much useful information, but it also creates information clusters here individuals overvalue investments in familiar equities or other investments. A filter or lens may be able to be implemented in systems which will mask names of companies and investments while focusing the decision maker’s attention on quantitative financial data such as P/E ratios. Making such tools available to investment decision makers may prove an effective way of promoting more rational investment decision making.
Risk appetite and management
An aspect of retirement income planning which tends to receive much less attention than it should is risk management. There are a broad range of risks which a rational retirement income decision maker would be expected to address. The most critical risks to be considered in retirement income planning include early retirement due to death or disablement, longevity risk, sequencing risk and market cycle investment risk. Like other decisions, the risk appetite of decision makers is potentially afflicted to behavioural biases which may result in irrational and inadequate decisions being made about managing retirement income risks.
Normalcy bias and optimism bias are two observable tendencies in decision making behaviour which have the genuine possibility of influencing decisions about the appropriate appetite for risk. Normalcy bias is the refusal to plan for a disaster which an individual has not experienced before. Retirement is an event which individuals will not generally experience until later in life. Retirement in itself is certainly not a disaster but rather something which should be valued; however, most people would consider retirement without access to income required for subsistence or a dignified retirement as a disaster.
The normalcy bias of expecting that an income will be available in retirement much as it is during working years has a significant influence over the decisions individuals about the level of savings. The identification of normalcy bias in the decision making of governments, fiduciaries and individuals may materialise as an understated rate of savings. In a similar tone, optimism bias is the tendency individuals have towards over optimism, overestimating favourable outcomes.
The illusion of control is the tendency for decision makers to overestimate the degree of influence over other external events. This scientifically observed trait may also result in a bias for individuals mistakenly feeling as though they are in control of certain risks over which they exert little or no control. This bias can see that individuals irrationally choose not to mitigate such risks through mutualisation or insurance of these risks, as they believe that they are in control of the risks of living longer than expected, early retirement or the impact of market cycles on investments.
While there is a quantifiable risk to individuals of early retirement due to disability, living longer than expected or the sequence of market movements negatively affecting retirement incomes, biases towards normalcy and optimism, coupled with the illusion of control may be contributing to irrational decisions being made by individuals and fiduciaries. Automatic acceptance into group insurance policies has proven to be relatively effective in broadening the coverage of insurance products designed to manage early retirement or disability risks for individuals exhibiting lower levels of engagement and involvement.
However, principles of risk profiling and management are often only applied at an enterprise level. When it comes to retirement income planning or management, there may be significant merit to developing mechanisms aimed at identifying the relevant risks to and controlling within an individual appropriate appetite.
Fees, costs & efficiency
It is impracticable for most individuals, and even fiduciaries to adequately implement a retirement income strategy without some outsourcing. Individuals rely on products and services supplied by third parties in implementing their retirement income planning strategy, hopefully appropriate to their level of engagement.
Naturally, this will see that fees and costs are incurred in exchange for the provision of these products and services. It’s important that retirement incomes aren’t adversely affected by excessive fees and costs which might be incurred by either the individual or fiduciary. Similarly, it is important to governments that systemic efficiency within the financial product and service supply chains.
The availability of information about fees and costs is an important characteristic of an efficient market, yet, there are also behavioural biases which can result in systemic inefficiencies in the markets for financial products and services which form the supply chains. Any market for retirement income focused financial products and services is unlikely to take a strong efficiency form due to the existence of behavioural biases.
The framing effect is a cognitive bias characterised by the arrival at different conclusions based on the same information due to the way in which it’s presented. The impact of the framing effect on decisions about fees and costs means that the disclosure of retail fees and costs to individuals, and wholesale fees and costs to fiduciaries can significantly impact decision making and therefore the efficiency of retirement income decisions and systems.
The presentation of fees as a percentage of a transaction or asset value are a classic example of this, with the presentation of the costs in a unit of currency resulting on the decision maker placing a greater weighting or emphasis on the competitiveness of the fees and costs. Similarly, the framing of fees as a percentage or in basis points disassociates the fees and costs from the currency required to pay them and results in a lower weighting and focus on the competitiveness of fees and costs.
Anchoring is a tendency for a decision maker to take existing fees and costs incurred for financial products and services and then apply it as a subjective reference point for making future decisions. This is particularly relevant with percentage based fees on a growing asset base. The supplier of a product or service which charges a percentage based fee will naturally be anchored at the current price.
The anchor of the current price will result in decision makers considering any reductions in the percentage based fee as a reduction, even where the asset base on which the fees are calculated in growing and therefore the actual amount of the fee in increasing. Asset based fees and costs are common in the supply of the financial products and services to retirement income systems. Ensuring that individuals and fiduciaries are provided information about the actual costs of financial products and services, including the cost at projected asset values.
Think again!
The ways in which the insights from behavioural finance and economics can translate into meaningful measures will differ significantly between jurisdictions, entities and individuals. Yet, there seem to be significant opportunities for behavioural science to provide insights which will prompt lawmakers and fiduciaries to think again. As changes to old age dependency ratios will present a daunting economic and political challenge to many western economies, improving the performance and efficiency of retirement income systems and pension funds is a challenge which will need to strike the right balance between individual liberty and objective paternalism.
