Skip to content
Total Portfolio Approach

Concept 1 of 6

Reference portfolio

A simple, cheap mix of listed shares and bonds that states the board's risk appetite and serves as the yardstick for every active decision.

Why this matters to you

Choose your view in the header to see why this matters to you and to start at the right level.

In plain words

A fund first writes down the plain, cheap mix of shares and bonds it could hold with almost no effort, its reference portfolio, which shows how much risk the owners accept. Everything else the managers do must beat that mix after costs.

Sources: CFA Institute Research & Policy Center 2026, p. 41; NZ Super Fund, Reference Portfolio page; A. Ang, M. Brandt, D. Denison for the Norwegian Ministry of Finance 2014, pp. 9, 22

Picture it

Picture two portfolios side by side. One is the real fund, with property, infrastructure and private companies. The other exists only on paper: the cheap mix the board chose to express its risk appetite.

Every way the real fund differs from the paper one is a deliberate choice by management. Each choice is, in effect, paid for by selling part of the paper portfolio, so the fund's market risk stays where the board set it. The gap between the two returns, after costs, is the value added.

Sources: NZ Super Fund, Reference Portfolio page; CFA Institute Research & Policy Center 2026, Exh. 4 p. 21; CalPERS Board 2025, slides 7–8; A. Ang, M. Brandt, D. Denison for the Norwegian Ministry of Finance 2014, pp. 9, 22

A €10bn fund and its reference portfolio

Illustrative numbers

The board of a €10bn fund picks a 70/30 reference portfolio: €7bn in a global share index and €3bn in a government bond index. Management builds the actual portfolio differently.

The paper portfolio and the real one

Both hold €10bn. The board's reference portfolio is a cheap 70/30 mix; the actual fund is management's choice, and every difference between the two is an active decision.

The paper portfolio and the real one
€bnListed sharesGovernment bondsPrivate real estateInfrastructure
Reference portfolio€7bn (€bn)€3bn (€bn)0 €bn (€bn)0 €bn (€bn)
Actual fund€5bn (€bn)€2bn (€bn)€2bn (€bn)€1bn (€bn)

Illustrative numbers

Actual holdings against the reference portfolio
HoldingActualReference portfolioActive decision
Listed shares€5bn€7bn−€2bn
Government bonds€2bn€3bn−€1bn
Private real estate€2bn€0+€2bn
Infrastructure€1bn€0+€1bn

Over the year the reference portfolio returns 6.0% and the fund returns 6.6% after all costs, so management has added 0.6 percentage points, or €60m. The board never set a property weight: the €2bn in real estate, and whatever it earns or loses, belongs wholly to management's record.

How funds do it

CalPERS is the simplest case. From July 2026 its board sets just two numbers: a 75/25 (as of 17 Nov 2025 · source) reference portfolio of global shares and US Treasuries, and an active-risk limit of 400bp (as of 17 Nov 2025 · source), capping how far results may stray from it. They replaced its strategic asset allocation, with its ranges and asset-class benchmarks. Its chief investment officer described the anchor as the portfolio the fund would build “if we only had a handful of people”. Any holding beyond the two indices, private assets included, is an active decision owned by staff.

  • NZ Super has used one since 2010. Its last published mix is 80/20 (as of Apr 2026 · source)1 growth assets to fixed income, fully hedged to the New Zealand dollar. The board chooses it; management decides how to invest against it. The fund uses it to set expected returns, judge results after costs and set hurdles for active investments.
  • CPP Investments now calls its anchors Market Risk Targets, formerly its Reference Portfolios. For the base plan the target is 85/15 (as of 31 Mar 2026 · source) global shares to Canadian government bonds, set above a minimum of about 60/40 (as of 31 Mar 2026 · source) that would just sustain the pension. The board approves them on the chief investment officer's recommendation. The Strategic Portfolio at the same risk holds 24% private equity, 24% real assets and −23% financing (as of FY2026 · source): the anchor fixes a level of risk, not a recipe.

Sources: CalPERS 2025, p. 1; CalPERS 2025, memo p. 2; deck slides 5, 7; CalPERS 2025, slides 5–6, 14–15; CalPERS 2025, transcript pp. 41, 43; Wilshire 2026, p. 2; CalPERS Board 2025, slide 10; NZ Super Fund, Reference Portfolio page; NZ Super Fund 2020, pp. 2–4; NZ Super Fund 2026, SPE p. 6; CPP Investments 2026, “Market Risk Targets”; CPP Investments 2026, pp. 14, 18–19, 88–89

How much in shares? Four funds' reference portfolios

Each board sets its fund's appetite for risk as a simple mix of shares and bonds. The bar shows the share part of the mix; the rest is bonds.

How much in shares? Four funds' reference portfolios
% in sharesShares in the mix
CPP Investments, base plan (Market Risk Target)85/15 (% in shares)
NZ Super Fund80/20 (% in shares)
CalPERS75/25 (% in shares)
GIC, April 2013 to March 202665/35 (% in shares)

Sources:

CPP Investments, base plan (Market Risk Target): 85/15, as of 31 Mar 2026, CPP Investments 2026, p. 18

NZ Super Fund: 80/20, as of Apr 2026, NZ Super Fund, Reference Portfolio page1

CalPERS: 75/25, as of 17 Nov 2025, CalPERS 2025, memo p. 2; slide 5

GIC, April 2013 to March 2026: 65/35, as of Aug 2013, GIC 2013, p. 212

Notes on the numbers

  1. 80/20, NZ Super reference portfolio, growth / fixed income: 75% global equities, 5% NZ equities, 20% global fixed income, 100% hedged to NZD: Last published mix. A new reference portfolio reportedly took effect on 1 Jul 2026, but its composition has not been disclosed (reported).
  2. 65/35, GIC Reference Portfolio, Apr 2013 to Mar 2026: global equities / global bonds: Historical. From 1 Apr 2026 GIC uses a Strategic Portfolio grouped by role (growth, income, inflation), with undisclosed weights.

Expert view and evidence

The CFA Institute's 2026 guide defines a reference portfolio as a simple mix, often listed shares and bonds, that expresses the board's risk appetite. It is one yardstick among several, not a portfolio to track closely: tracking error against it, the typical yearly gap between the returns, is usually around 4–5% a year (as of Jul 2026 · source). The idea was formalised in the 2014 review of Norway's fund, whose Opportunity Cost Model judges every deviation, net of fees, against the securities that pay for it.

Choosing the mix is about risk preference, not optimisation: NZ Super's three 2015 candidates all had a Sharpe ratio of 0.22 (as of 2015 · source). Designs differ. GIC replaced its 65/35 (as of Aug 2013 · source)1 anchor in April 2026 with a role-based Strategic Portfolio, CPP no longer scores itself against its targets, and the Future Fund has none, steering by an equivalent equity exposure of 65 (as of Aug 2025 · source). For the CFA Research Foundation, an anchor is a matter of governance design, not a requirement of the total portfolio approach (TPA).

The evidence is thin. No controlled study shows that TPA funds, with or without an anchor, beat a traditional strategic asset allocation, and the CFA Research Foundation's 2026 review finds the literature does not establish systematic outperformance. Single-fund records point both ways: NZ Super earned 9.68% (as of 30 Jun 2026 · source)2 a year over the 20 years to June 2026, against 8.19% (as of 30 Jun 2026 · source) for its reference portfolio, while CPP earned 8.8% (as of 31 Mar 2026 · source) a year over ten years, against 10.7% (as of 31 Mar 2026 · source)3 for its Market Risk Targets.

Sources: CFA Institute Research & Policy Center 2026, pp. 20–21, 41; A. Ang, M. Brandt, D. Denison for the Norwegian Ministry of Finance 2014, pp. 9, 22; NZ Super Fund, 2015 Review App. 8 p. 30; FY2026 results release; GIC 2013, p. 21; GIC 2026, Box 3, pp. 20–21; CPP Investments 2026, pp. 20, 34, 43; Investment Magazine 2026, P4–P5; Future Fund Board of Guardians 2025, para 5.13; CFA Institute Research Foundation 2026, pp. 8, 39

Notes on the numbers

  1. 65/35, GIC Reference Portfolio, Apr 2013 to Mar 2026: global equities / global bonds: Historical. From 1 Apr 2026 GIC uses a Strategic Portfolio grouped by role (growth, income, inflation), with undisclosed weights.
  2. 9.68%, NZ Super Fund annual return, 20 years to 30 Jun 2026: Self-reported record of one fund, not a controlled comparison.
  3. 10.7%, CPP Investments Market Risk Targets annual return, ten years to 31 Mar 2026: Since FY2025 CPP no longer measures its relative performance against these targets, which it describes as an expression of targeted market risk rather than an investable alternative.

Common misunderstandings

“The reference portfolio is the mix the fund is trying to hold.”

In fact:

It is a yardstick, not a target mix. NZ Super states that its reference portfolio is a benchmark and not a guide to what the fund actually holds, and CPP Investments carries its target's level of risk with large private holdings and borrowing that a two-index mix does not have.

Sources: NZ Super Fund 2020, p. 2; CPP Investments 2026, p. 19

“Something that simple must be easy to beat.”

In fact:

Not reliably. NZ Super returned 14.17% (as of 30 Jun 2026 · source) in the year to June 2026, just behind 14.27% (as of 30 Jun 2026 · source) for its reference portfolio. In a stylised model from CPP Investments, even a diversified design expected to add value has a 29.8% (as of Jul 2026 · source)1 chance of trailing its benchmark over ten years: a model probability, not a record of what happened.

Sources: NZ Super Fund, FY2026 results release; CPP Investments Insights Institute 2026, pp. 4, 12

Notes on the numbers

  1. 29.8%, CPP Investments stylised model: probability that a diversified design expected to add 1% a year (61% correlation with its benchmark, 880bp tracking error) trails that benchmark over ten years: A stylised probability from CPP's model with assumed inputs (jointly normal returns, 10% volatility), not a backtest or a historical frequency. CPP says neither design in its illustration is inherently superior.

“Every fund that uses TPA has a reference portfolio.”

In fact:

Many do, but it is not required. Australia's Future Fund says it deliberately has none, and the CFA Institute's guide describes the reference portfolio as a common feature of TPA rather than a necessary one.

Sources: Investment Magazine 2026, P4–P5; CFA Institute Research & Policy Center 2026, p. 20

“The reference portfolio is the fund's return target.”

In fact:

Not on its own. The CFA Institute's guide says it expresses risk appetite and gives each deal a notional source of funding, but does not by itself set the return the fund is aiming for.

Sources: CFA Institute Research & Policy Center 2026, Exh. 4 p. 21

Check yourself

CalPERS' board has set a 75/25 (as of 17 Nov 2025 · source) reference portfolio and a 400bp (as of 17 Nov 2025 · source) active-risk limit. Management buys a stake in a private real estate fund. Which statement is correct?

Related

Sources

  1. The Total Portfolio Approach (TPA): A Practical Guide for Navigating the Transition to TPA. CFA Institute Research & Policy Center (R. Urwin & G. Hayman), 2026.
  2. Total Portfolio Approach: A Critical Literature Review. CFA Institute Research Foundation (R. Elkamhi and J.S.H. Lee), 2026.
  3. Review of the Active Management of the Norwegian Government Pension Fund Global. A. Ang, M. Brandt, D. Denison for the Norwegian Ministry of Finance, 2014.
  4. Our Investment Strategy (Total Portfolio Approach, Market Risk Targets, Strategic and Benchmark Portfolios). CPP Investments, 2026.
  5. Measuring What Matters: Evaluating the Total Portfolio Approach. CPP Investments Insights Institute (S. Shen, D. Walker, G. Rubin), 2026.
  6. Statement of Investment Objectives, Policies, Return Expectations and Risk. CPP Investments, 2025.
  7. Annual Report 2026 (fiscal year to 31 March 2026). CPP Investments, 2026.
  8. Reference Portfolio. NZ Super Fund (Guardians), web page.
  9. How We Invest white paper: The 2020 Reference Portfolio Review. NZ Super Fund, 2020.
  10. Statement of Performance Expectations 2026/27. NZ Super Fund, 2026.
  11. NZ Super readies reference, tops the world, down on global shares. Investment News NZ, 2026.
  12. Statement of Investment Policies. Future Fund Board of Guardians, 2025.
  13. 'Every opportunity has to earn its place': How the Future Fund built a TPA culture that scales. Investment Magazine (L. Maddock), 2026.
  14. Report on the Management of the Government's Portfolio 2025/26. GIC, 2026.
  15. GIC's New Investment Framework. GIC, 2013.
  16. ALM: First Reading of PERF Recommendations (staff deck). CalPERS (S. Gilmore, M. Nix, S. Terando), 2025.
  17. ALM: PERF Recommendations (Investment Committee agenda item 5a). CalPERS, 2025.
  18. Board Education Day: TPA and transition considerations. CalPERS Board, 2025.
  19. Total Fund Policy Review: First Reading (Wilshire memo). Wilshire (T. Toth), 2026.
  20. Investment Committee transcript (TPA adoption meeting). CalPERS, 2025.
  21. ALM: Reference Portfolio Design and Active Risk Limits. CalPERS, 2025.