Investigating behavioural biases of pension scheme trustees

The importance of decisions made by pension trustees

As the population ages, it is widely accepted now that state pensions alone cannot cover the living costs of the future retired population.

To help with this issue, governments have been encouraging individuals to seek additional pension income in the form of private pensions. According to the UK Pensions Commission, individuals cannot expect the state pension to cover even half of their ideal pension income, in particular those with higher earnings, with the vast majority of pension income required expected to come from private pensions.

As participation in private pensions increases worldwide, the value of assets in pension funds continues to grow. The UK Pensions Regulator estimates that 84% of UK staff have enrolled in a private pension scheme in 2018, up from 40-50% about 10 years before. The Organisation for Economic Co-operation and Development (OECD) estimates that in 2018 pension funds controlled US$28 trillion in assets in the OECD countries, equivalent to 53% of their GDP. With such a large presence, pension funds yield substantial power which can influence the markets.

As most private pension schemes are managed by boards of trustees of five to seven people, this power is in the hands of a few individuals. Their decisions can move markets and influence the retirement outcomes of millions of pension scheme members. Yet, very little research has looked at how they make decisions – and if any biases could be leading to suboptimal outcomes.

In the old world of Defined Benefits pensions, poorly made investment decisions (by the trustees) would place additional burden on the sponsor (typically the employer) to compensate for any deficits. Now, in the new world of Defined Contribution pensions, if trustees make poor decisions, it directly affects the pensions of scheme members.

In summary, bad investment decisions can have catastrophic consequences towards the pensions of future retired individuals. Dr Iain Clacher and I were engaged by the The Institute and Faculty of Actuaries (IFoA), partnering with researchers from City, University of London; Ipsos; University of East London and Reverence Limited, to research behavioural biases in the decision making of pension scheme trustees, with the aim of improving the decisions made by trustees, and ensuring better outcomes for the members.

Behavioural finance biases

As we know from previous research, individuals are very poor at making rational investment decisions. There are many behavioural finance biases that can lead to poorly performing investments.

For example, investors typically display a bias towards naïve diversification bias. When asked to distribute investments across many options, individuals tend to allocate money evenly across all of the options available. If there are two options (such as two funds or two stocks), they will allocate half of the money to each, and if there are three options, one third will be allocated to each, and so on.

Individuals’ choices are then determined by the number of options available, regardless of the underlying nature of the options, and without considering the individual characteristics and merits of each. This can easily lead to suboptimal asset allocation, and is also open to exploitation by a third-party who can manipulate the list of potential options available.

Another relevant bias which can translate to suboptimal investment returns is the disposition effect. Investors are typically reluctant to sell (even large) losses, and eager to quickly realise (even small) gains. This can translate into assets being bought high and sold low. It can also lead to excessive trading, which can be costly in terms of brokerage commissions and fees.

However, all of the above research relates to individuals making decisions with their own money. When members pay contributions into a pension scheme, they outsource most of the investment decisions to the trustees of pension schemes. Even though scheme members are typically able to choose their own pension parameters eg whether to invest in the high-risk or low-risk option, the vast majority of individuals will opt to stay with the default option, which is set by the scheme trustees.

Almost no previous academic research has been dedicated to studying the decisions made by pension scheme trustees before. What little research there is shows trustees fall for the same behavioural biases as everyone else, but these studies examined personal financial decisions rather than the institutional investments that trustees make on behalf of scheme members. The remit of our current project is to empirically test the behavioural finance biases that might affect the investment decisions of trustees.

The setting in which trustees make decisions

Since there is very limited research exploring the decisions made by trustees, we analysed research conducted with other individuals (not trustees) but in similar settings to those in which trustees make decisions. We know that trustees rely heavily on external advice (such as financial and legal consultants), they make surrogate decisions on behalf of others (the members), and they make decisions in groups. Our in-depth review of these areas can be found here. Below, we summarise our findings.

Advice: The Myners report (a 2001 report for HM Treasury, reviewing institutional investment in the UK) on pensions was particularly concerned about external advice provided to trustees by experts such as investment consultants. It reported that the advisers could be exerting excessive influence on the decisions made by trustees.

Past research on advice-taking has shown that individuals are usually receptive to advice, but that they tend to put more weight on their own judgments, reducing the influence of advice. However, many factors can increase the weights given to advice, making it more influential.

Many of these factors are present in pension trustee decision-making. For example: when individuals pay for advice, when a decision is complicated, or when individuals need to diffuse responsibility (legal liability of trustees). These factors could be putting unwanted decision control in the hand of external advisers.

Surrogates: Surrogates (individuals who make decisions on behalf of others) tend to make very poor decisions for others. When making surrogate decisions, individuals tend to project their own preferences, instead of considering the preference of the other. This can be an issue when the surrogate and other person are different, which tends to be the case with trustees and members: trustees are often predominantly male, older, richer, better educated, and very often already retired.

While the gold standard for surrogate decisions is to choose the action exactly the same as the other would choose, most often surrogates tend to choose the action that the other should choose. Therefore the surrogate tends not to honour the choices of the other party if they disagree with them. Surrogate decisions are also emotionally more muted, rational, and less empathic, which makes it difficult for them to truly relate to the needs of the others.  

Groups: Group decisions are not efficient: they are plagued by process losses. Members of brainstorming groups will typically wait for others to come up with ideas, censor their own ideas (believing them to be bad), and take ownership of ideas that were not their own. Group discussions rarely create information sharing between the group members, with group decisions tending to be based on information that was previously shared anyway, before the group met and discussed. This leads to lower productivity per person and lower number of ideas being generated in groups than in individual decision-making.

Group discussions also tend to lead to choice-shift and group polarization, with more extreme decisions at group level than at the individual level. As a result, trustees might not be pooling expertise, and might be making inefficient group decisions, and extreme decisions such as too much (or too little) risk taking.

Conclusion: As far as we are aware, no behavioural research has empirically tested pension fund trustees’ decisions to investigate how the combination of group decisions, external advice, and surrogacy influence their decisions and, ultimately, the sustainability of our pensions. Given how much influence trustees’ decisions have on asset allocation and, by extension, in financial markets, this is a surprising state of affairs.

New empirical research is underway

After analysing the existing research on trustee decision making, we created a series of experiments in order to collect behavioural data directly from pension fund trustees and test the potential impact of some of the biases identified above. Access to trustees was provided by our partner, Aon UK. The objective of the data collection process is to empirically identify any biases, and engage with policy makers on how to circumvent such biases and improve the decision-making process of trustees, ensuring a positive outcome to future pensions in the UK.