Home > News > 2004 > April 2004 > Accounting for Pension Costs under FRS17 - 2003 Survey
Accounting for Pension Costs under FRS17 - 2003 Survey
Barnett Waddingham's report on the financial assumptions used by the FTSE100 companies under FRS17.
Introduction
This is our third annual survey of the assumptions adopted by FTSE100 companies for determining the value of their pension liabilities under the Financial Reporting Standard, FRS17.
Our 2003 survey is based on the data in the published accounts of 42 FTSE100 companies with financial years that ended on 31 December 2003. All of those companies included were also contained within our 2002 survey of FTSE100 companies.
Discount Rate
The discount rates used by the companies in our survey are set out below.
At 31 December 2003, the average yield on over 15 year, AA-rated corporate bonds was approximately 5.4%, unchanged from the previous year-end.
In our 2002 survey only 2 companies used a discount rate at or below the average rate, but in 2003 this has increased to 22. As well as the general downward shift in the discount rate, there has been a narrowing of the range of discount rates used: in our 2002 survey, the discount rates ranged from 5.2% up to 6%, whereas for 2003 the discount rates used fell in the narrow range of 5.25% to 5.6% as shown in the table above.
In our 2002 survey, we suggested that companies had slightly increased the discount rate relative to the average AA bond yield in response to the pressure put on FRS17 deficits by the fall in asset values during 2002. Likewise, there would now appear to be some evidence that companies have responded to the improvements in asset values during 2003 by adopting discount rates closer to the average AA bond yield. Another effect that we have experienced is that the choice of assumptions is coming under greater scrutiny from the auditors who are increasingly challenging the use of discount rates that fall too far away from the average AA bond yield.
Inflation Rate
FRS17 states that the relative yields of index-linked and fixed interest bonds of similar credit standing will give an indication of the expected rate of general inflation. As at 31 December 2003, this difference in yields was approximately 2.8%. The chart below confirms that this was the most popular choice of assumption, although the majority of companies in our survey used a lower rate.
Salary Increases
Some companies may use a scale for promotional salary increases in addition to a general salary growth assumption and therefore a comparison of the disclosed salary increase rates may not be a like-for-like comparison in all cases. Nevertheless we have shown below the disclosed salary increase assumption relative to the assumed rate of price inflation (i.e. the real rate of salary increases assumed).
Surplus/Deficit
The increase in equity asset values during 2003 generally improved FRS17 deficits, although this was offset to some degree by a reduction in discount rates and the increase in assumed inflation. Overall, the average FRS17 funding level for the companies in our surveys improved from approximately 80% as at 31 December 2002 to approximately 84% as at 31 December 2003.
Expected Investment Returns
The expected long term investment returns disclosed as at 31 December 2003 are used in the calculation of the P&L pension costs for 2004. The higher the expected returns, the lower the 2004 P&L pension cost. The expected returns on equities disclosed at 31 December 2003 are set out below.
The above data reflect an average Equity Risk Premium (defined as the excess of expected equity return over the gilt yield) of 3.1% per annum, compared to 3.5% as at 31 December 2002.
FRS17 states that the amount of the expected return on assets is expected to be reasonably stable. Given the substantial return on equities during 2003, one would expect the percentage expected return to be reduced at 31 December 2003. Indeed most companies either reduced or left the equity return assumption unchanged from 2002, however a few companies chose to increase the assumption, as illustrated in the chart below.
The assumed expected return from bonds was less variable between companies, as shown below.
Asset Allocation
The chart below compares the percentage of assets invested in equities for the companies in the survey at 31 December 2003 with the percentage weighting at 31 December 2002. The straight line on the chart represents the position of an unchanged equity weighting from 2002 to 2003.
The chart shows that the proportion of assets invested in equities remained relatively unchanged from 2002 to 2003. Some companies appear to have allowed the equity weighting to drift upwards slightly, reflecting the strong returns on equities relative to other asset classes during 2003, whereas other companies appear to have re-balanced to keep close to or slightly below the 2002 equity weighting. There is only one company in our survey that significantly reduced its equity weighting during 2003, but perhaps more will follow suit during 2004.
Mortality Assumption
Another assumption that can have a significant impact on the FRS17 results is the choice of the mortality basis, i.e. the assumed life expectancy of the members. If the projected improvements in life expectancy are underestimated, and/or if an out-of-date mortality analysis is being used, a company could be significantly understating the value of its FRS17 liabilities. For example, the impact on the FRS17 deficit of a change to the latest published mortality tables from the previous tables (i.e. PMA80 to PMA92) could be equivalent to a change of around 0.5% per annum in the discount rate. Including allowance for the recently published cohort projections could be equivalent to a further discount rate change of 0.25%. Hence, we find it surprising that FRS17 does not require the disclosure of the mortality assumption used, and we would expect auditors to give the mortality assumption as much attention as the financial assumptions
Summary
The survey shows that although there are areas of consistency, many variations still exist between assumptions used for reporting under FRS17.
However, it is particularly interesting to see that companies are increasingly using a discount rate very close to the average yield on AA corporate bonds and that the divergences seen in previous years have larger been eliminated. This may reflect an increasing interest being shown by the auditors in the choice of assumptions as we approach the date when companies will have to fully adopt FRS17 (i.e. for financial periods starting on or after 1 January 2005).
We hope that this analysis is helpful to companies in formulating their own FRS17 assumptions.
Please contact your usual Barnett Waddingham consultant to discuss any of the above.
Barnett Waddingham, April 2004.