PBU works out to 'get fit'
The Pensionskassen for Børne- og Ungdomspædagoger (PBU) scheme has more than 72,000 members with premium benefits of DKK1.3bn (e175m) a year. Before the end of 2000, PBU had been wholly administered by Sampension. At that time, PBU had been imbued with a defensive attitude, costs that were deemed far too high and a limited investment policy that impacted the fund’s yield.
In 2001, the scheme embarked on a ‘fitness project’, a set of objectives, which comprised of four parts: to achieve a high and value-assured yield in the long term, to reduce administration costs, to set new payment patterns and to improve dialogue with its members.
The road to achieving a high and value-assured yield began with PBU taking a conscious risk management approach to allow the purchase of more volatile assets, such as equities.
To do this, PBU decided to divide itself into two independent bonus and investment departments. The old department now accounts for benefits based upon a fundamental interest rate of 4.25% and the assets that are associated with the benefits. The new department accounts for the benefits that are based upon the other basic rates in PBU (ie, 1.6% and 2.6%) with associated assets.
PBU now has an investment policy that can be optimised in relation to the individual department’s basic interest rate. In the old department, the main emphasis is placed upon ensuring that the yield matches the profitability requirements from the reserves. In the new department, the objective is to secure a high real rate of return, whereby the purchasing power of the pensions is maintained after inflation.
As a result, the shareholding in the new department has increased from 8% to a current level of 32%. PBU’s ability to withstand a fall in interest rates has been improved and will be further improved by the setting up of a third department without a yield guarantee.
PBU’s new administrative model has led to it appointing ATP PensionService to manage the member administration and accounts functions; Nordea as the principal cooperation partner regarding investment; and Kuben to take care of the property administration.
The new model has led to significant annual cost savings. The costs percentage, which had previously amounted to about 5-6%, dropped to 3.6% in 2001.
Regarding new payment patterns, the general fall in the basic interest rate meant that, as a result of the special composition of PBU’s benefits pattern, new members were not achieving the intended risk cover. PBU has now worked out a new benefits pattern that affords everybody the requisite risk cover, and will be implemented at the beginning of 2003. This will enable members to move their current arrangements over to the new department with the new – and more appropriate – benefits pattern.
Finally, in order to improve dialogue with its members, PBU has ensured that its general assembly consists of delegates who are elected locally by the members. In recent years, the delegates have been increasingly included in the general and fundamental decisions regarding the scheme, and management information is now sent to them on a quarterly basis. In 2002, PBU established a web site, www.pbu.dk, for its members, and introduced membership meetings.
PBU has also been employing value at risk (VAR) as an internally established risk objective, which means the chosen asset composition within the board’s laid down frameworks are not too risky when viewed in relation to the estimated buffer.
Recognising the need to concentrate on the risk of interest falls, PBU has developed the concept of ‘Value Loss’ in co-operation with investment manager Nordea as an expression of the increase in the allocations that the estimated interest fall may involve. This has been important as the allocations in the old department are around double the duration of long-term government bonds.
VAR is unique insofar as it is calculated on the basis of a one-year horizon. The buffer is also budgeted for a year at a time in order to compare with the calculated VAR. The risk tolerance is set at 2.5%. Calculation of correlations and standard deviations are based upon monthly time series from 1996–2002.