In the context of retirement security, technology now makes it relatively inexpensive to provide individuals with a wide range of financial choices, and the platform costs of providing choices will certainly decline further in the future. Nevertheless, there is a broad literature indicating that, when people are faced with disparate and complex choices, they do not necessarily make better decisions. At the same time, it may be difficult to determine the wisdom of individuals’ decisions until a future date, as individual circumstances vary and there is much noise in outcomes.
Many approaches are available to ensure better decisions in the retirement saving and dissaving context. First, people often need help making choices, and the use of default options and other behavioural approaches can enhance these decisions. These can produce more uniform decisions, and for some, may provide protection against downside risk. Yet such approaches can also fail to properly account for diversity in preferences and circumstances. Moreover, individuals offered default options are less likely to learn about financial alternatives and unlikely to argue for better retirement savings products.
Second, individual retirement saving decisions can improve if people can get better financial advice. Nevertheless, existing advice channels have weaknesses: financial planners are sometimes conflicted, and employer-provided distribution networks may not be as cost-effective as they should be. Moreover, a variety of approaches have been tried, ranging from price regulation on products, to the regulation of distribution networks, to disclosure requirements, to the limitation of commissions. No single approach on its own seems to produce unambiguously better outcomes.
A third approach would be to train people so that they can make better retirement saving and decumulation decisions on their own, taking greater individual responsibility for their own financial wellbeing. Yet the problems here are also manifold since financial product complexity can make it difficult to determine even what people need to learn. Mitchell et al. (2017) do confirm substantial welfare improvements when workers are provided financial education. Still, employers are often reluctant to take responsibility for workplace-related training and advice. Moreover, more research is required to pinpoint the cost-benefit tradeoffs for the provision of financial literacy.
This special issue of the Journal of Pension Economics and Finance offers new and exciting evidence on this third aspect of retirement preparedness. The paper by Barua, Koh, and Mitchell (Reference Barua, Koh and Mitchell2018) looks at a natural experiment in Singapore where college students were provided financial education. The study found that pupils taking a financial education course had significantly better knowledge and outcomes. The authors are aware of potential endogeneity concerns since students electing to study financial literacy could be more financially literate already. To correct for this possibility, the authors use the fact that students bid to enrol in the course, but not all bidders were selected. Accordingly, the authors could compare students who just missed making a successful bid with those who successfully bid to take part in the course. Their difference-in-difference approach showed significant gains in financial knowledge and literacy among those who took the course. Whether it is worth more or less than other alternatives is hard to tell: financial literacy helps people minimize risks and ensure better planning, whereas other courses could enhance labor market outcomes.
The paper by Giesecke and Yang (Reference Giesecke and Yang2017) asks whether greater financial knowledge and literacy helps people make smarter pension system choices. In particular, pension system incentives often have weak effects, so this paper explores the hypothesis that this could be due to participants’ incomplete financial knowledge. The authors use a multiperiod laboratory experiment with 318 individuals starting close to retirement, who were provided with different levels of knowledge about their circumstances. Results show a substantial difference in retirement decisions based on the level of knowledge, with a mean difference in retirement age of close to 2 years between the control and treatment group. Though the sample size is small, the case study does point to potentially quite high rates of return for more knowledgeable individuals at older ages. Whereas the implication of work by Barua et al. (Reference Barua, Koh and Mitchell2018) could be that expending money informing younger individuals would be helpful, the results in Giesecke and Yang (Reference Giesecke and Yang2017) indicate substantial benefits of improving older individuals’ understanding.
The Joulfaian (Reference Joulfaian2018) paper uses the US Internal Revenue Service panel data to look at retirement savings decisions by the self-employed; this subset of the population naturally has more retirement flexibility than do other workers. One reason individuals might not retire is that they may need to first pay off debt, and the author explores this proposition. He concludes that the self-employed use the mortgage tax deduction in the USA as a means of saving more for retirement. Yet business debt operates differently than does mortgage debt, so overall it reduces peoples’ saving. It would be interesting in the future to evaluate how these results interact with financial knowledge in the self-employed population.
Kalmi and Ruuskanen (Reference Kalmi and Ruuskanen2018) focus on financial literacy and retirement planning in Finland, where the retirement system is among the more complex, with relatively high levels of state and statutory benefits. In general, the population is highly literate and scores well on metrics of financial literacy. The authors conducted 1477 face-to-face interviews and reports substantial inequality of financial literacy across the population. Interestingly, the self-employed were among the best informed. The study found no direct link between basic metrics of retirement planning and financial literacy, overall, though even here, women fared less well.
The work by Keim and Mitchell (Reference Keim and Mitchell2018) assessed what happened when the choice set of options for a defined contribution plan was halved, though the range of fund options remained relatively similar. In this case, workers whose fund choices were eliminated either moved to other funds or were defaulted into a target date fund. Using administrative data, the authors showed that trimming options led to less churn across funds and lower risk. They then calculated the 20-year cost savings for the individuals who were streamlined, and it proved to be quite material. The main explanation for the cost savings is that the plan sponsor retained the lower-fee funds while eliminating the higher-cost funds.
The final paper in our collection, by Bockweg et al. (Reference Bockweg, Steenbeek and Vonken2018), looks at framing and annuitisation in a Dutch pension fund. In this case study, retirees could allocate up to 20% of their accruals to a lump sum, but the decision was framed differently for different participants. This paper is particularly interesting since much of the literature on retirement plan framing has focused on US cases. While the institutional context and level of annuitisation in the Netherlands differs from elsewhere, the authors show that the extent of framing is still quite high. One exception is that individuals with high debt levels prefer a larger lump sum, compared with others holding less debt.
The papers in this special issue contribute insights that can help improve pension designs and decisions in the future. Moreover, as technology makes providing choice less expensive, research in the area must focus more on how to make sure that the choices provided are understood clearly and structured in a manner to add the most value.
