Listen
In 1994, Swedish parliament legislated a completely new pension system based on a defined contribution design. The new system was implemented in 1999. The driving force behind the reform was the realisation that the overall design of the then reigning defined benefit ATP pension scheme was not resilient to fundamental economic and demographic shocks. So, in 1999 the existing flat-rate folkpension, which was topped up by the defined benefit earnings-related ATP scheme, was replaced by mandatory non-financial defined contribution (NDC) and financial defined contribution (FDC) schemes, with a new minimum pension benefit that is means-tested against the NDC and FDC pensions.
Together the NDC and FDC schemes work as lifecycle savings accounts in which the individual account balances are converted into a universal longevity insurance with the same rules for all. Individual account values and life expectancy at retirement determine the yearly pension amount for the remainder of life. The overall design makes it possible to retain a constant contribution rate over generations, strengthens the incentive to work longer, and maintains financial sustainability.
The success of the NDC and FDC schemes in achieving intergenerational fairness and long-run financial sustainability relies on projections of life expectancy that neither systematically under- or overestimate life expectancy over a succession of birth cohort “pools” of pensioners. It is only when this statistical criterion is fulfilled that financial sustainability and fairness over generations can be maintained.
Examining birth cohort data for the period 1900-2014 for eight OECD countries (Denmark, France, Italy, Netherlands, Norway, Sweden, the United Kingdom and the Unites States) and employing the two most popular projection models of official agencies and national pension schemes – the “period method” and the Lee-Carter model – the study shows that this criterion is not fulfilled over time for any of the countries. The study shows that this is because these two methods depart from an assumption that the rate of change in mortality is a constant value, whereas the empirical evidence is that the overall picture is one of steady improvement in the rate of change.
Authors
Edward Palmer, Professor Emeritus of economics, Uppsala University
Yuwei Zhao de Gosson de Varennes, PhD in economics, Uppsala University, currently Analyst at the Swedish Tax Agency.
This report is part of the research program “New Challenges for the Pension System”. The program takes an overall perspective on the pension system and deals with questions that concern both the occupational pension system and the state pension.