Will Equity Investment Improve Pension Funding?
By: Doug Andrews
Where do equities fit into a pension plan’s investment policies today? Can they enhance returns, improve the funded position of a plan, and reduce contribution requirements? Doug Andrews, of Aon Consulting, examines the Equity Risk Premium.
For more than 40 years, most pension plan sponsors and administrators have looked to equities as an asset class to enhance overall returns. At the beginning of this millennium, equity markets took a nose dive – sharply reducing the funded ratio of pension plans as they fell. Now, we wonder if, in the next 10 years, equities can be expected to regain their status as an asset class capable of enhancing returns, improving the funded position, and reducing contribution requirements.
There are many answers to this question. This article briefly reviews reasons for both positive and negative answers, but does not reach a conclusion. It is intended to give pension plan sponsors and administrators information as they consider how to invest the pension plan’s assets and what the longer term contribution requirements will be. It provides ammunition for sponsors and administrators as investment managers and actuaries are put on the firing line to justify their return expectations and financial projections.
What Is The Equity Risk Premium Assumption?
The funded position of a pension plan is determined by comparing the assets to the liabilities. But in an ongoing plan, at any point in time, the liability value is an estimate. Its ultimate value will depend on such factors as the mortality, morbidity, and termination experience of the membership; benefit levels granted; and the investment returns achieved. To place a value on liabilities at a single point in time, the actuary makes a number of assumptions. A significant assumption affecting the estimate of the liabilities is the investment return assumption.
Within the last five years, there has been increasing attention directed to the choice of the investment return assumption. Proponents of the financial economics approach to assumption-setting have argued that Defined Benefit pension promises are similar to the issuance of a bond from the plan to the members. To ensure security of the promised benefits, the plan should value the liabilities with expected bond-like returns.
Let us accept this logic, assume that liability values have been estimated using a bond-like interest rate, and turn to the asset component of the funded ratio. If there were assets equal in value to the liability estimate, and if those assets were invested to earn the assumed bond-like return, today’s liabilities would remain fully funded (ignoring any deviations from other assumptions). As a starting point in understanding future potential contribution requirements, sponsors and administrators should ask, ‘what are contributions expected to be if a 100 per cent bond investment philosophy is used?’
From this starting point, it is logical to ask, ‘can the assets be invested to earn more than the expected bond-like returns?’ Traditional wisdom is yes. Because there is additional risk assumed by an equity investor relative to a bond investor, (because the equity investor has a lower order entitlement to a company’s revenue), there must be an equity risk premium (ERP). In other words, if an investor could choose to invest in the equity or the debt of a company, the investor would only choose to invest in the equity if the investor expected to receive a higher return for receiving less security.
If we accept this argument (and there are some who would object), what is the expected ERP? This is an important question for the sponsor and administrator to consider if they pursue an investment policy including equities (which is by far the most common investment approach).
A Range Of Equity Risk Premiums
So what is an appropriate ERP? There is a range of answers. Aon Consulting gathers information from more than 25 of Canada’s major institutional investment management firms, twice yearly, regarding their forecasts of returns on the major financial market indices for various time horizons, ranging from one to 10 years. The managers do not explicitly forecast the ERP. However, by comparing their forecasts of expected returns on the equity indices compared to the bond index, one observes a positive ERP which is implicit in their forecasts. Table 1 shows a range of implicit equity risk premiums using the forecast returns at January 1 for the years 2001, 2002, 2003, and 2004.
Over a 10-year horizon, the implicit ERP is 2.2 per cent to four per cent. Conversations with investment managers suggest that a range for the long-term ERP of two per cent to four per cent is commonly quoted.
An examination of Canadian economic statistics compiled by the Canadian Institute of Actuaries indicates that the implied ERP for the 50-year period ending December 31, 2003, is approximately 3.2 per cent, comfortably in the range suggested above. However, as we examine these statistics for shorter periods, such as 10 years, we note a much lower ERP in the more recent 10 year periods. For the 10-year period ending 2003, the ERP was approximately 0.1 per cent.
Robert D. Arnott and Peter L. Bernstein have carried out research and written several papers regarding ERP 1. Although their research focuses on the U.S. financial markets, Arnott has suggested that similar results could be expected in Canada. The U.S. longer term history (50 years) shows an ERP of approximately five per cent. Their analysis suggests that expected real stock returns over the 10 year period ending 2012 will be 2.5 per cent, which is less than expected real bond yields of 3.3 per cent and means a negative ERP of 0.8 per cent.
Bernstein writes, “the primary reason that the actualized equity risk premium has been so enormous in the years since 1954 is that the bond market went through repeated inflationary episodes that took bond investors by surprise” 2. A similar explanation may apply to the CIA data cited previously. Few investors expect inflationary episodes in the next 10 years.
What Should Sponsors And Administrators Do?
Where one believes the ERP will fall – in the range from four per cent to negative 0.8 per cent – will significantly affect assetmix decisions.
Bernstein suggests portfolio policies of “50/50 [equities and bonds] or no policy numbers at all, purely an opportunistic strategy based on no fixed opinions – at least until equity valuations change enough to reveal which asset (stocks or bonds) will be the better performer in the long run” 3.
This author suggests the following approach for sponsors and administrators considering an asset allocation policy:
- Understand the value of liabilities using bond-like rates.
- In making asset-liability projections, use as a starting point a 100 per cent bond allocation. The characteristics of the bond portfolio should ideally match the characteristics of the liabilities as closely as possible, particularly in areas such as duration, inflation sensitivity, and cash flows.
- When considering other asset allocations, understand the implicit ERP.
- Test a variety of ERP assumptions to understand the potential variability of results and the impact on financial position.
- If you are not satisfied with the results, do further analyses considering alternative asset classes.
Doug Andrews is a vice-president at Aon Consulting in Toronto.
1. See, for example, Peter L. Bernstein, ‘Determining the Equity Risk Premium,’ AIMR Equity Research and Valuation Techniques, (2002) or Robert D. Arnott and Peter L. Bernstein, ‘What Risk Premium is ‘Normal’?’ Financial Analysts Journal, (March/April 2002).
2. Peter L. Bernstein, op. cit.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -