Pension Investing – A Brief History and Future Questions
Authored By: Jeremy Bell | Publish Date: June 26, 2018
Authored By: Jeremy Bell | Publish Date: June 26, 2018
In this article, I discuss the changes in pension investing, particularly related to the relationship between pension sponsors and pension investment consultants. My intention in this article is to provide a rough sketch of the last 20 years to provide a perspective on current investment policies, processes and habits related to pension plan investments.
The ultimate goal of the article is to challenge plan sponsors to consider how they carry out their investment function going forward. At the end of this article, I suggest a handful of specific areas that plan sponsors might consider.
My thinking is informed by my work experience, nearly 20 years in Western Canada. As a general rule, I have found:
Finally, I have used specific years in this article when talking about the history of the industry. I felt that specific years assisted the narrative, although I recognize that different sponsors, firms and geographies evolved at different speeds. I recognize, for example, that some jumbo pension plans (like Ontario Teachers Pension Plan and Canada Pension Plan Investment Board ) evolved earlier and past the scope of this article.
Investment Consulting – Early Years: Pre-2007
In the 1990s, large actuarial consulting firms in Canada created standalone investment consulting departments. Prior to that point, investment advice was generally either not provided by anyone, or provided by plan sponsors’ actuaries in addition to their roles calculating liabilities and service costs as well as provide general ad hoc consulting related to governance, regulations and plan design.
Investment advice was initially provided on the fee model in place for actuarial consulting, fee-for-service at hourly rates.
Investment options were fairly limited in this period: fewer options were readily available to most pension plans and even if more were available, pension sponsors would have been unlikely to dramatically depart from 60%/40% equity/fixed income allocations . Investment consulting was largely limited to:
Pension sponsors that lengthened bond duration benefited from the advice of their investment consultant  in hindsight. Similarly, adding real estate was generally beneficial in hindsight based on the level and stability of returns earned.
The primary activity of investment consultants during this period was the search for traditional investment manager (i.e., public stocks and bonds) skill:
The Benefits of Investment Manager Research for Traditional Asset Classes
That deep investment manager research in traditional asset classes would be expected to result in skilled managers and better outcomes over time is an attractive concept. An industry exists that charges money to sophisticated investors for this service: why would sophisticated investors choose to purchase that service if it didn’t add value?
Unfortunately, the logic and evidence in support of value from manager research on traditional asset classes is weak:
Given the inherent difficulty of the task, the existence of sufficient data within large consulting firms to perform analyses (and to support ongoing analyses) and the enormous financial benefits associated with providing evidence of this ability, one should be very skeptical that manager research can provide benefits in traditional asset classes .
Plan sponsors must consider this lack of evidence when considering what investment committees spend their time on and what they pay for providers to do on their behalf.
Investment Consulting – Alternative Assets: 2007-Present
In the next phase of investment consulting, the search for added value (alpha) in traditional asset classes was largely set aside. The manager structure and habits in place to support this search largely remained in place (the amount of committee time and money spent on monitoring), but manager termination decisions became less frequent and, where managers were terminated there was more consideration of passive options.
Sponsors and consultants did not give up on the value of manager research providing added value in traditional asset classes, instead their efforts were redirected toward introducing alternative asset classes to portfolios. Many sponsors would have started this period with 0-10% of assets (perhaps real estate and/or another asset class) in alternative asset classes and would have ended this period with 10-30% in alternative asset classes.
Alternative asset classes were introduced with the aim of providing some combination of bona fide diversification from stocks, interest rates and credit (the dominant risks in most pension funds) and attractive returns for a given level or risk .
By the end of this period, pension plans were generally making better use of their advantages (scale and long time horizons – required for some, but not all alternative asset classes) and were slightly more resilient (i.e., able to handle very poor economic environments). For many pension plans, however, the changes have been incremental. A deep, prolonged economic contraction combined with an increased pace of corporate competition would put significant pressure on pension plans: public stocks, corporate bonds, private equity and real estate would all be affected to varying degrees.
Investment Consulting – Today Forward
There are currently two distinct investment implementation paths:
There is no correct answer to this choice. In fact, if delegated investment managers present external, rigorous and ongoing evidence of added value relative to the current state, smaller consulting firms will present this option to their clients. Today, however, it would be difficult to recommend this option.
The implementation options (i.e. whether to engage investment managers directly or indirectly) are less important than the asset allocation going forward. Plan sponsors should be thinking about the following:
The desire to turn investment decisions into solvable math problems is attractive. Making decisions using the resulting probability distributions from asset allocation studies is much easier than the alternative. But, models provide a false level of precision. Even more complicated models still provide a false level of precision.
Most asset allocation studies attempt to limit the influence of the model on decisions, through setting constraints on asset mixes in an optimization process (i.e., no more than 5% of assets can be invested in emerging market equities in any modeled portfolio) and through slightly departing from the modeled asset mix(es) when selecting a policy asset mix. These actions implicitly accept that the model is (and must be) an imprecise approximation of the real world and only to be trusted to guide incremental decisions.
As a result, these studies are generally only helpful with making incremental decisions, often fairly infrequently. Plan sponsors and their consultants should spend more time thinking about the stresses and shocks that their portfolios might face and the opportunities available. A more frequent, more open minded discussion (not just an update of the existing model) on the asset mix would be a useful supplement to asset allocation studies as they generally occur today.
From a pension sponsor’s perspective, if their real estate fund is borrowing money at a cost of 3.5% per year and the pension fund loans money through a bond fund at 2.5% per year, the total of the two transactions is a loss of 1% of the amount of loans . Unless there is some aesthetic reason or unique circumstance to maintain a certain level of bonds, this sponsor should sell bonds and pay down its mortgages, to the extent possible, or if a lower leverage real estate fund is available, the sponsor should sell bonds and move the real estate investment to the lower leverage fund.
Similarly, where levered and unlevered funds are available from a manager, sponsors should seriously consider making a larger allocation to the unlevered fund than would be made to the levered fund, either funded by selling bonds or by efficient borrowing by the pension plan.
Plan sponsors should actively consider their advantages and consider whether they could responsibly make better use of these advantages.
Plan sponsors and pension investment consultants have evolved a fair amount over the past 20 years. Even still, both could go further in breaking the habits of pension investing to allow sponsors to wrestle with interesting, consequential questions going forward.
Jeremy Bell FCIA, CFA
Partner at George & Bell Consulting
 Persistent manager skill in many asset classes should be rare (if a manager has a “secret recipe” such that they can consistently add value after fees over time, they will eventually be replicated by others, decreasing added value opportunities for a given level of risk).
 There are other jumbo pension plans or jumbo public sector pension investment managers whose evolution and current state is reasonably described in this article.
 There was limited need to consider other asset classes: bonds provided decent returns, say 5-7% per year, roughly sufficient to fund a pension promise based on prevalent assumptions.
 Investment consultants were not prescient in predicting decreases in interest rates, but instead understood that pension liabilities are essentially negative long-term bonds – most apparent when a solvency test applies, but true regardless. The best way to relieve the risk of a negative long-term bond is to buy a long-term bond. Interestingly, investment managers were generally unsupportive of extending pension plans’ bond duration, and some managers fiercely objected.
 Ironically, these consulting firms insist on subjecting external managers to external, rigorous and ongoing scrutiny.
 What good is changing investment managers quickly when there is no evidence that the new highly-rated manager will perform any better than the now out-of-favour manager? Equivalently, if a drug is not shown through evidence to have clinical benefits, why bother increasing the dosage?
 There are plenty of anecdotes that could claim to show evidence of this ability. Anecdotes of added value from traditional asset class manager research are as compelling as a gambler’s anecdotes of winning bets as proof of gambling ability: not at all.
 Volatility is a shorthand for risk and is often used interchangeably. This is reasonable for most, but not all, purposes. For example: an investment with a certain annual return of 3% should not be considered less risky than an investment with an annual return that will alternate between 4% and 5%. The second investment is more volatile but is certainly less risky by any reasonable standard.
 The entire banking system is built on the concept of taking liquid deposits and lending those assets on an illiquid basis.
 Putting this in a personal context: if you have a $100,000 mortgage at a 4% interest rate and you have $100,000 invested in bonds (say outside of a registered plan for simplicity) yielding 2.5%, you should sell the bonds and pay down your mortgage. You still own the same house, but now you won’t be losing $1,500 per year from a peculiar financing decision (probably worse because you would likely pay tax on the bond return).