Description of the programme

The university provides retirement benefits to its employees through a traditional mandatory pension plan but employees can also make voluntary contributions to a complementary Tax Deferred Account (TDA) 403(b) plan. Each year, the university organises a benefits fair where all employees are invited to come and learn about the different kinds of benefits (including retirement benefits) provided by the university.

At the same time, every employee receives enrolment forms. However, employees are free to enrol in the TDA or change their contribution level or investment decision at any time throughout the year.

The study

The aim of the study was to explore the effect of information and social interaction within the university on enrolment in a retirement plan.

The experiment encouraged a random sample of employees in a subset of departments to attend the benefits information fair organised by the university, by promising a monetary reward ($20) for attendance. It then investigated the subsequent enrolment in the TDA at an individual and departmental level. There are two distinct treatments in the experiment: receiving the letter, and being in the same department as someone who receives a letter.

The sample was restricted to staff employees (i.e., non-faculty employees) aged less than 65 and eligible to participate in the TDA but not enrolled (6,200 individuals).

The average number of employees per department is 30, but the median size is much smaller, around 15. Therefore, except in a few large departments, it is expected that each employee knows most of their colleagues in the department.

What are the outcomes?

• the study assessed the difference in enrolment rates between those receiving an incentive to attend the benefits fair; those in a department as someone who receives an incentive and those in a department where now one receives an incentive
• the study assessed the difference between attendance at the benefits fair for the groups above

Key findings

• TDA enrolment at 5 and 11 months after the fair was significantly higher in departments where some individuals were incentivised to attend than in departments where nobody was treated. However, there is no significant difference in TDA enrollment between those who actually received the encouragement letter and those in the same departments who did not. Being in a treated department increases TDA enrolment rate by 0.93 and 1.25 percentage points (after 4.5 and 11 months).
• The increase in TDA contribution generated by this experiment was much larger than its costs. The incentive increased participation after one year by about 1.25 percentage points, generating an estimated additional TDA savings of about $200,000 per year compared to a cost of the rewards distributed of about $12,000.
• The large effect of a small reward on fair attendance, amplified by social effects, suggests that individuals do not optimally seek out and process available information on their own. This suggests that individuals may not be giving much thought to their retirement savings decisions, which has important policy implications.
• Small financial incentives have successfully encouraged treated employees, as well as members of their peer groups, to attend a benefits fair. Treated individuals (i.e. receiving financial incentive) are more than five times as likely to attend the fair as control individuals (28% with standard error 0.6% compared to 5% with standard error 0.5%). Non-treated individuals in treated departments are three times as likely to attend the fair as control individuals in non-treated departments, despite the fact that only original letter recipients could claim the $20 reward. This shows that the invitation letters not only increased the fair attendance rate for individuals who received them but also had a spill-over social effect on their colleagues within departments.

Points to consider

• • Strong randomised trial design, incorporating factors such as departmental size
  • The experiment does not allow for separation of differing social effects but does conclude that social interaction plays an important role in decision-making
• • Of interest to those wishing to increase the proportion of individuals contributing to pension plans within the workplace
• • The social interaction effect may not necessarily be transferable to a different context, for example one with less coherent teams or organisations where employees are under greater work pressure which might impact on attendance at a benefit fair