The Rise of the Reference Portfolio
Building benchmarks to reflect changing pension assets.
December 8, 2011
Move over policy portfolio. Increasingly, institutions are moving to reference portfolios. That move reflects a shift from a passive indexed portfolio to an actively managed one that is more and more invested in private market assets.
The Canada Pension Plan Investment Board is now in the second version of its reference portfolio in the past five years says John Ilkiw, currently an independent consultant and a former senior vice-president CPPIB. Established in 1997, its initial objective was a 65%-equity/35%-bond split, and as a start-up, CPPIB invested passively.
He made his remarks this month at a presentation on portfolio benchmarks sponsored by the Canadian chapter of the Alternative Investment Management Association (AIMA) in Toronto.
CPPIB’s strategies – indeed, its investment objectives — have changed over the years, hence the need for a reference portfolio. The portfolio breaks out asset categories, and provides a low-cost, all-index and broadly diversified allocation that seeks to capture systematic risk – or the market return. But the reference portfolio is not the actual investment portfolio.
CPPIB’s board determines the reference portfolio; its active managers have broad flexibility in implementing it. The idea is to relieve the board of getting into highly complex investment structures they weren’t expected to understand but nevertheless had to approve. In addition, the reference portfolio sets out the active risk measures; thus the investment managers have a annualized risk measure of 90% Value at Risk – meaning normally no more than a 10% drawdown. (“Normally” is a freighted word, however, that hides more than it discloses when it comes to financial markets under duress.) The only time the board is involved in daily investment decisions is when they concern very large transactions, or when they might jeopardize CPPIB’s reputation.
The reference portfolio provides a benchmark that is “easy to understand, easy to explain and easy to measure,” Ilkiw explains. It also frees the hand of investment managers. The current reference portfolio is 45% foreign developed market equities, 25% Canadian nominal fixed income, 15% Canadian equities, 5% Canadian real return bonds, 5% emerging market equities, and 5% foreign sovereign bonds.
What the investment managers add are two things. The first is the “better beta” component, namely asset classes that are not yet amenable to being passively indexed, for example private equity or real estate. Then there is the quest for “incremental risk-adjusted returns” through active management. Often, says Ilkiw, these two cannot be separated and he cites private equity investing as an example: leverage and alpha both contribute to returns.
But that creates a risk to be managed. CPPIB runs a “giant long/short portfolio,” Ilkiw adds. Since private equity is still equity, to add 10% exposure (CPPIB is currently running exposure at 15%), CPPIB would have to sell off 10% of its public equities to stay within the reference portfolio.
But it’s more complex than that, since private equity is normally leveraged by about a third. Thus, to be risk neutral, a further portion of the public equity portfolio is sold to fund the fixed income portfolio to keep the effective equity exposure constant.
When it comes to judging investment success, Ilkiw suggests a two-by-two matrix with two variables: market risk and active risk. If markets perform as expected, and if the active management program also performs, it’s a success — that’s the upper left-hand or northwest quadrant. But there are instances in the matrix where one or the other – or both – don’t perform. Market risk – Ilkiw labels that “an Act of God” — has not been rewarded lately, while active management has contributed 17bps to total portfolio returns. The matrix allows for accountability to the CPPIB board.
Like CPPIB, University of Toronto Asset Management has adopted a reference portfolio approach. Unlike, CPPIB, however, it is not a direct investor but rather a “manager of managers,” says Bill Moriarity, UTAM’s CEO. UTAM has almost 100 investment managers.
It has followed the U.S. endowment model, pioneered by Yale and Harvard. There are a couple of considerations behind that model. The first, given the general ebullience of equity markets plus declining inflation “you didn’t have to be a rocket scientist to generate great returns” in the 1980s and 1990s, Moriarity explains. But things changed after the bursting of the Technology-Media-Telecommunications bubble in 2000. Before that investors could count on a traditional 60/40 portfolio invested mostly in domestic stocks and bonds to deliver around 10% real return in the U.S. and 7% in Canada. Over the past decade, that traditional portfolio has delivered around 2.4%– much less than the 5% endowments need to sustain funding commitments and maintain purchasing power.
As a consequence, endowments have increasingly shifted to a “multi-asset diversified portfolio,” which downplays domestic securities in favour of foreign and private market ones, and moreover, overweights equities in virtue of their potential contribution to funds that exist as perpetual pools of assets.
But once one moves beyond traditional domestic assets, a strong investment infrastructure is required, he says. And, a more sophisticated staff understanding of how asset allocation works. Indeed, 2008 proved tumultuous for UTAM. “It’s no secret that the University of Toronto had a difficult time in 2008.” He attributes that to a focus on asset class returns rather than the underlying risks of the portfolio. UofT wasn’t alone: many portfolios were “well-positioned for normal times, but not for times of substantial stress.” But UofT also had many private equity investments where it had neither the scale to get lower fees, nor the expertise to evaluate the risks, Moriarity says. “We just didn’t have the resources to invest effectively.”
As it turned out, many apparently different asset classes – hedge funds and private equity — shared common underlying return drivers with other elements of the portfolio, and that has led to a different approach at UTAM. The first was the adoption of a reference portfolio, meant to be a “shadow portfolio” that would be appropriate to a university endowment. That resulted in a asset allocation of 30% to U.S. equities, 15% Canadian, 15% EAFE, 35% fixed income and 5% real return bonds, with currency 50% hedged.
The second was to divide exposures into five different risk buckets: economic growth, credit risk, interest rate risk, inflation risk and currency risk.
But that’s not a paint-by-numbers approach, Moriarity says. He points out that Harvard and Yale, following a similar endowment model, have very different actual asset allocations. Instead, the reference portfolio is a process that has a “fourfold benefit:” better understanding of the return drivers; better understanding of how the portfolio is likely to perform; a better basis for portfolio discussion;and a better understanding of potential stresses.
That eventuates in “better beta,” as a “more granular approach to risk metrics” helps educate stakeholders who, on a behavioural basis, tend to be pro-cyclical.
And that, he feels is necessary to get beyond what is on offer in the market: a 2.5% real return.