“Take into account the model uncertainty associated with the scenario sets”
What is the focus of the paper?
Pension funds use uniform scenario sets to calculate future pension benefits. This paper examines how (un)stable these so-called ‘P- and Q-sets’ from DNB (the Dutch Central Bank) are from quarter to quarter. It then explores the causes and consequences of these variations and discusses possible ways to better handle the changing scenario sets.
What are the key findings?
Significant changes can occur in the uniform scenario sets from quarter to quarter. These changes have a substantial impact on the various metrics used by pension providers. The researchers conclude that the variations in successive scenario sets can often only be partially attributed to objectively observable economic developments. To an extent, the changing scenario sets, and the widely varying pension outcomes based on them, are the result of model and parameter uncertainty. The researchers discuss four ways to address this: (1) abolishing the uniform scenario sets; (2) fixing the long-term model parameters; (3) publishing multiple versions of the scenario sets (based on different model and parameter choices); and (4) accounting for model uncertainty when using the scenario sets. These solutions are addressed separately in the paper for the P-set and Q-set.
What are the implications?
- The use of uniform scenario sets must be handled with care: the danger is that the outcomes are seen as the absolute truth.
- It is important to emphasize in communications with pension participants that the presented pension amounts involve uncertainties and are not guaranteed.
- Given the expanding use of the scenario sets in the Future of Pensions Act, an evaluation of all these
applications is recommended.