DB: What went wrong?

Robin Pond | May 11, 2011

Perhaps it’s premature to do a post-mortem on corporate DB plans, as there are still plans in existence. And, perhaps the DB plan will arise—like a phoenix—transformed into some more sustainable target benefit form.

But it is useful to ask, ‘What went wrong?’ since the answer calls into question some fundamental assumptions that many plan sponsors continue to rely on when setting investment policies, assumptions such as the following:

1. The equity risk premium is positive.
2. A pension fund is a long-term investor and therefore should have a long-term investment policy.
3. Given the long-term nature of pension investments, the fund’s asset mix targets can be expressed in terms of static percentages.

In theory, these assumptions make sense. In practice, they have turned out to be false.

Risk premium
One of the basic tenets of investment theory is a positive relationship between risk and return. An investor is compensated for bearing risk. This assumption underlies much of our pension investment and funding policy.

In a standard asset/liability model this is embodied as an equity risk premium, an assumed long-run return premium for stocks over bonds or cash. When setting a going concern discount rate, an actuary may award a fund with more stocks versus bonds a higher discount rate on the assumption of higher future investment returns.

But this is not supported by the risk/reward data for various indices in Canada for a 30-year period up to the end of last year.

For the past thirty years in Canada, there has been no reward for bearing equity risk. Long bonds have outperformed stocks over this period, and using long bonds as a rough proxy for the return on pension liabilities, it is clear why the growth in pension assets has failed to keep pace with the growth in liabilities.

Sitting in 2011 with 20/20 hindsight, it is tempting to ask why plans did not jump at the chance to lock in the high long bond yields back around 1981. If they had, much of the funding pain they are currently experiencing could have been avoided.

But interest rates were extremely high in 1981 for a reason. Inflation was out of control and locking into fixed nominal rates, at any level, was considered extremely risky.

Also, while a move to longer bonds today might be considered de-risking in relation to solvency and marked-to-market accounting liabilities, it would increase funding volatility in relation to a relatively fixed going concern discount rate. Ironically, it was the decline in interest rates that made solvency the dominant funding consideration for most plans, but this only happened in the later 1990s when solvency discount rates started to dip below the going concern assumptions.

Long-term investment
Connected to the assumption of a positive risk premium is the belief that pension fund investments are long-term, so we should ignore shorter-term noise and manage to a long-term policy.

John Maynard Keynes, in his oft-quoted passage, questions this belief: The long run is a misleading guide to current affairs. In the long run we are all dead. Economists set themselves too easy, too useless a task if in tempestuous seasons they can only tell us that when the storm is past the ocean is flat again.

When stocks underperform bonds for 30 years, even pension funds—the classic long-term investors—may not be able to ride out a shorter-term in which economic realities stray significantly from economic assumptions. In the long run, a lot of DB pension funds may be dead, just because the long run is proving to be too long, and the assumptions too incorrect.

Static asset mix
The long-term assumption led to establishing static asset mix targets and managing to them as if everything including economic conditions, plan demographics, funding levels, and sponsor circumstances were unchanging. In retrospect, an asset mix policy that adjusted more easily to changing asset valuations, changing funding levels, and changing sponsor risk tolerances would have made more sense. An important part of setting a plan’s asset mix is defining the parameters that should cause the mix to change.

Asset mix shifts should not just be driven by our attempts to outguess the markets. Policy asset mix for most pension funds shifted considerably over the past thirty years. Back in the 1980’s a typical plan might have had 60% in fixed income and 40% in equities. By the end of the 1990s this flipped to around 60% in equities and 40% in fixed income.

The on-going shift to equities was motivated by a lingering fear of inflation, a prevailing belief that interest rates were just about to rise, and unwavering faith in the ability of equities to outperform in the long run. Again, in retrospect, more attention to risk tolerances as opposed to market expectations would have yielded a better result.

It is an open question whether current moves into alternative investments and de-risking strategies in reaction to a low interest-rate environment may be any more successful than the policy moves of the past 30 years. The best answer remains a very clear definition of objectives, an examination of risk tolerances and a parameterized policy to dynamically manage to these objectives and tolerances.