Framing and pension annuities: Experimental evidence from a Dutch pension fund

This thesis has been awarded the Netspar Master Thesis Award 2015, sponsored by the UWV Pension Fund.

Recently a debate in the Netherlands has emerged on an increasing demand for individualized pension solutions and the implementation of partial lump sum cash outs in second pillar pension plans. This thesis is a first assessment of how Dutch plan members will react to such new optionalities. In a survey-based experiment, about 3000 members of a Dutch occupational pension plan were asked to allocate their real projected pension accrual between a life annuity and a partial lump sum. In the experiment most respondents made use of this new optionality. In a second step, the driving variables of annuity and lump sum demand are analyzed. Departing from standard economic theory as well as human biases in decision taking, the predominantly behavioral nature of the decision to annuitize is revealed. Particularly strong evidence for the framing hypothesis in annuity demand is presented. Dutch plan members are found to filter the underlying decision through a “consumption frame” rather than an investment frame. Individual decisions appear to be steerable when actively framing plan members. Next to framing, default setting is assessed as a tool to actively steer individual decision taking. The third step of analysis refines insights from the framing hypothesis: Robust evidence is found for the reaction to active framing being subject to individual characteristics, highlighting the impact of heterogeneity on framing effects. The thesis concludes, that while plan members welcome the partial lump sum option, pension professionals must be aware of the impact they have on overall outcome through information architecture. Framing and default setting are found to be powerful “nudges”, capable of predictively steering annuity demand.