What companies listed on the S&P/TSX Index do in regard to the accounting standards applicable to employee future benefitsLinkedIn
There, in black and white
NB Bulletin Vol. 8 N. 16, December 2005
Several companies in the private sector will soon be finalizing their financial statements for the fiscal year that is drawing to a close. Among the disclosure notes that should be prepared for these financial statements one pertains to costs for defined benefit (DB) pension plans and group insurance plans offered to retirees (other post-employment benefits).
The required assumptions to determine the accrued benefit obligation and the cost must be established by the management of the company. To help companies select these assumptions, we have analyzed the annual reports of 26 Quebec-based companies listed on the S&P/TSX Index whose fiscal year ended between September 30, 2004 and March 31, 2005. All of these companies sponsored at least one DB pension plan and 22 of them offered at least one group insurance plan to their retired employees.
Our analysis is divided into three sections. The first section presents the assumptions used in percentile format. The charts present a comparison of the assumptions used for fiscal years ending in 2004 (between September 30, 2004 and March 31, 2005) and in 2003 (between September 30, 2003 and March 31, 2004). The second section summarizes the results of these charts along with additional data analysis. The third and final section presents our comments with respect to the selection of some assumptions.
1.Selection of assumptions
The following charts present the survey results of the key economic assumptions used for DB pension plans and other post-employment benefits in percentile format.
| ||End of 2004|
(Sept. 30, 2004 to March 31, 2005)
|End of 2003|
(Sept. 30, 2003 to March 31, 2004)
|Defined Benefit Pension Plans|
Expected Long-Term Rate of Return on Assets
Compensation Growth Rate (also applicable in some cases to other post-employment benefits)
|Other Post-Employment Benefits|
Trend Rate for Health Care Costs - Initial Period
Trend Rate for Health Care Costs - Ultimate Period
- For DB pension plans, the median discount rate used at the end of 2004 was 6.00%, compared to a median rate of 6.25% at the end of 2003. Close to 80% of the surveyed companies used a discount rate between 5.75% and 6.25% at the end of 2004. As a result, there is little variance in this assumption. Although the median discount rate fell by 0.25% from 2003 to 2004, about one quarter of the companies maintained or increased their discount rate between those dates. The observed range in this assumption might be explained, in particular, by the demographic profile of the members of the surveyed companies and the relative weight of DB pension plans outside Canada (a different discount rate may be used for DB pension plans outside Canada).
- The median expected long-term rate of return on DB pension plan assets was 7.50% for 2004. All but one of the surveyed companies used an assumption equal to or higher than 7.00%. Since actuarial valuations may be performed once every three years, it is conceivable that this assumption might be reduced at the next actuarial valuation, as it was the case for many funding valuations made as of December 31, 2004 or more recently. Although the purposes of funding valuations might be different from those of accounting valuations, a discount rate below 7.00% was almost always used for our clients’ funding valuations as at December 31, 2004.
- In accordance with the new disclosure requirements of the Canadian Institute of Chartered Accountants (CICA), at the end of 2004, companies were required to identify the proportion of DB pension plan assets invested in equities and fixed income securities. Close to 80% of the surveyed companies had between 50% and 60% of DB pension plan assets invested in equities. As the proportion of assets invested in equities should be aligned with the investment policy, the range in the expected long-term rate of return assumption was not explained entirely by the actual asset mix. However, similarly to the discount rate, the relative weight of DB pension plans outside Canada (a different expected long-term rate of return may be used for DB pension plans outside Canada) may partially explain the variance in the expected long-term rate of return assumption.
- Many companies sponsor DB pension plans and group insurance plans to their retirees that offer salary-based benefits. These companies must establish an assumption related to the compensation growth rate. At the end of 2004, the median compensation growth rate was 3.50%, a rate identical to the one used at the end of 2003. Half of the surveyed companies used a rate that varied between 3.50% and 4.00% at the end of 2004. The differences can be explained, in particular, by a company’s business sector, demographics of the covered group and the definition of compensation.
- For companies offering a group insurance plan to their retirees, the median discount rate was identical to the rate used for DB pension plans. At the end of 2004, around 30% of the surveyed companies were using a discount rate that was higher than the rate disclosed for DB pension plans, which might be explained by the demographic profile of the members (they may not be the same) and the timing of benefits (incidence and level of benefits may be different).
- The trend rate for health care costs is the most important assumption for group insurance plans offered to retirees. Given the significant increases witnessed over the last few years, health care costs are generally expected to rise at a higher rate in the short term (initial period) and gradually fall to a lower ultimate level. About 60% of the surveyed companies had used an initial rate higher than 9.00% in 2004. The variation of the trend rate might be explained, in particular, by the nature (drugs, hospital, dental care, etc.) and the particular experience of the plans.
- The ultimate median trend rate for health care costs was 5.00% at the end of 2004, compared to a rate of 5.30% at the end of 2003. Around two thirds of the companies used an ultimate rate between 4.50% and 5.50% at the end of 2004. These rates likely vary for the same reasons as the initial trend rates vary.
The discount rate selected should generally be established based on the rates of return for high quality debt instruments with cash flows matching the timing of benefits.
In the United States, the Securities and Exchange Commission (SEC) has stated that high quality debt instruments correspond, for example, to a Moody’s AA corporate bond rating. The CICA has not provided any opinion on the specific level of high quality debt instruments. The CICA did, however, suggested that an appropriate measure was the return on a long term Government of Canada bond with an adjustment to account for the risk premium applicable to high quality corporate bonds.
In Canada, given the scarcity of long-term bonds that meet the American standards for high quality, some Canadian companies interpreted the notion of high quality as an A rating or higher, or even a BBB rating in some instances. This more liberal interpretation generally resulted in companies using a discount rate 0.25% higher than a rate selected according to the American standards. The use of a discount rate reduced by 0.25% would have translated into an increase of approximately 5% to 10% in the cost of employee future benefits.
Moreover, the CICA with the intervention of the Auditing and Assurance Standards (AASB) issued on November 15, 2005 an invitation to comment on its proposed new standard-setting approach that envisions the convergence of the Canadian standards with the International standards. The International standards require, for the determination of the discount rate, the use of high quality corporate bonds. However, when there is a scarcity of high quality corporate bonds in a local market, the International standards suggest using a government bond rate of return. In this context, it is conceivable that a lower discount rate would have been used by some companies.
Finally, other factors, like the materiality of results or the extrapolation method used for long-term maturities, may explain the variance in discount rates used by the surveyed companies.
- Expected long-term rate of return on assets
As stated in Section 2, more recent valuations may prompt management to review and reduce the assumption related to the expected long-term rate of return on DB pension plan assets. Based on the information presented by the surveyed companies, we believe that some companies were overestimating this assumption. Decreasing the expected long-term rate of return assumption would significantly increase the cost of employee future benefits.
Please feel free to contact us for additional information.
630, René-Lévesque Blvd. West, 30th floor
Montreal, Quebec, H3B 1S6