What does CPP Investments hold?
At 31 March 2026, CPP Investments' net investments of $793.8bn were 35.6% public equities, 21.6% private equities, 20.2% real assets, 13.4% government bonds and 9.2% credit, by our calculation. The weights come from the dollar amounts in its ten-year review; the annual report rounds them to 36%, 22%, 20%, 13% and 9%.
At 31 March 2026: public equities 35.6% ($282.9bn); private equities 21.6% ($171.3bn); real assets 20.2% ($160.4bn); government bonds 13.4% ($106.4bn); credit 9.2% ($72.8bn). Source: CPP Investments Annual Report 2026, PDF pp. 36, 79 and 148; chart data (CSV). The weights are our calculation from net investments by asset class.
"Net" matters here. The weights above use the published net investment amounts from the ten-year review, including $106.4bn for government bonds. Gross investments of $1,015.9bn are a different measure from the $793.8bn net allocation; do not apply the net weights to the gross total. Borrowing and other financing exposures must be examined separately from those weights.
The Fund is two accounts. The base CPP account held $712.9bn and earned 8.0% in fiscal 2026; the additional CPP account, started in 2019, held $80.4bn and earned 5.4% with a lower-risk mix (PDF p. 38). By 30 June 2026, the combined Fund had grown to $863.6bn, after a 7.5% net return in the quarter and $10.1bn of net transfers from the Canada Pension Plan (Q1 fiscal 2027 release).
Where does CPP Investments invest?
The United States was 48% of the Fund at 31 March 2026 and 51% at 30 June 2026; Canada was 12% at both dates. Asia Pacific was 18% and then 17%, Europe 17% and then 16%, and Latin America 5% and then 4%.
31 March 2026: United States 48%, Asia Pacific 18%, Europe 17%, Canada 12%, Latin America 5%. 30 June 2026: 51%, 17%, 16%, 12%, 4%. Source: Annual Report 2026, PDF p. 36 and CPP Investments, Our Performance (retrieved 3 November 2026); chart data (CSV).
The Fund keeps much of that foreign exposure unhedged. CPP Investments says fully hedging back to Canadian dollars would not improve its long-term risk and return, and that it hedges selectively to hold a target Canadian-dollar exposure (PDF p. 22). That choice cost it in fiscal 2026: a weaker US dollar produced a foreign-currency loss of $12.4bn, a 1.8-point drag on the year's return (PDF p. 37). In the following quarter, a stronger US dollar added to returns.
What is the 85/15 Reference Portfolio, and is it still the benchmark?
The 85/15 mix is the base account's risk target: the market risk of 85% global public equities and 15% Canadian government bonds, held since fiscal 2019 with limits of 80% to 90% equity. CPP Investments now calls it the Market Risk Target and estimates the minimum risk the base account needs at 60/40; the additional account targets 55/45 against a 45/55 minimum (PDF p. 20; PDF p. 133).
The equity share of the target has risen from 65% in fiscal 2015 (55% global and 10% Canadian equities) to 85% from fiscal 2019.
The 2016 annual report called the same two-asset construct the Reference Portfolio. It records that in 2006 CPP Investments made "the strategic decision to move away from largely index-based investments towards the more active selection of investments", and set up the Reference Portfolio "as the passive investment benchmark" (Annual Report 2016, PDF p. 25).
Since fiscal 2025, the 85/15 mix is no longer the yardstick for performance. "In fiscal 2025, the Benchmark Portfolios replaced the Market Risk Targets as our benchmark against which relative performance is measured" (Annual Report 2026, PDF p. 22). The Benchmark Portfolios combine the public-market index assigned to each investment strategy, weighted by that strategy's target size. Articles that still describe an 85/15 index benchmark are describing the earlier framework.
Between the risk target and the holdings sits a long-term Strategic Portfolio. For the base account in fiscal 2026 it was 23% public equity, 24% private equity, 40% public fixed income, 12% credit and 24% real assets, with minus 23% in cash and absolute-return strategies, a negative weight that by our reading reflects the borrowing that funds the rest (PDF p. 21). Its intended regions were 11% Canada, 76% other developed markets and 13% emerging markets.
The Fund's net return was 7.8% in the year to 31 March 2026, 6.6% a year over five years and 8.8% a year over ten, and 7.5% in the quarter to 30 June 2026. Net income was $56.9bn in fiscal 2026 and $60.2bn in the June quarter, which CPP Investments calls its largest quarterly net income on record (Q1 fiscal 2027 results, p. 5). The ten-year figure rose to 9.4% a year to June 2026.
Fiscal years to March: F2007 12.7%; F2008 −0.4%; F2009 −18.8%; F2010 14.7%; F2011 11.6%; F2012 6.3%; F2013 9.8%; F2014 16.1%; F2015 18.3%; F2016 3.4%; F2017 11.8%; F2018 11.5%; F2019 8.9%; F2020 3.1%; F2021 20.4%; F2022 6.8%; F2023 1.3%; F2024 8.0%; F2025 9.3%; F2026 7.8%. Source: Annual Report 2016, PDF p. 130 and Annual Report 2026, PDF p. 148; chart data (CSV). The earlier decade is net of operating expenses and the later decade net of all expenses, so the two halves are not on an identical basis.
Two of the twenty years lost money, both around the 2008 financial crisis. On fiscal-year-end values, the Fund fell about 19% from March 2007 to March 2009 and was back above its March 2007 level by March 2011, by our calculation. Fiscal years end in March, so these figures miss the low points between year-ends.
Against its own benchmark, the record is mixed. CPP Investments reports net value added of 0.7% a year over ten years and 0.1% over five, but in fiscal 2026 the Fund trailed its Benchmark Portfolios by 5.4 points (PDF p. 36). It attributes the gap mainly to index returns concentrated in large US technology and communication-services companies, and to strength in power and energy, while "the Fund's private market allocations also participated less in these market dynamics than public market benchmarks" (PDF p. 37).
How much does CPP Investments borrow, and why?
At 31 March 2026, recourse leverage was 33.0% of net assets for the base account against a Board limit of 45%, and 20.3% for the additional account against a limit of 30%. In dollars, recourse leverage totalled $252.0bn, plus $3.3bn of limited-recourse debt in holding subsidiaries (PDF p. 138). For the base account, it rose from 23% of net assets ($122bn) in fiscal 2022 to 33% ($236bn) in fiscal 2026 (PDF p. 50).
CPP Investments gives two reasons. Leverage lets it diversify more, particularly through larger fixed-income holdings, and raise its gross holdings of every asset class while staying at its targeted risk (PDF p. 21).
In other words, it replaces some listed shares with private companies and some government bonds with credit, then borrows so that the whole still carries 85/15-level risk with more holdings underneath. Borrowing to hold more of the lower-risk assets is also the idea behind Ray Dalio's All Weather strategy and AQR's risk parity work, although CPP Investments combines it with large private holdings.
The borrowing has a cost. In fiscal 2026, financing expenses were $7.4bn, alongside $1.8bn of operating expenses and $5.9bn of investment-related expenses, mostly external managers' fees embedded in funds (PDF p. 52). The operating expense ratio alone was 23.1 basis points (PDF p. 5); GPIF's management and custodian fees, a narrower measure, were 0.02% of assets. CPP Investments reports its returns after all of these costs.
Do private assets make CPP's returns look safer than they are?
Private assets change when losses and gains show up in reported returns: CPP Investments wrote in 2016 that "there is typically a valuation lag between public and private markets". The full passage says "Reported private asset returns tend to be less volatile than comparable public market returns and there is typically a valuation lag between public and private markets" (Annual Report 2016, PDF p. 50).
It added that in sharp public-market losses it would expect to outperform the Reference Portfolio because of its private assets, "just as we would expect little or even negative DVA in periods of sharp public market gains". Fiscal 2026, with a 5.4-point shortfall in a concentrated public-market rally, fits the second half of that description.
The values themselves are estimates. At 31 March 2026, $442.8bn of the $793.8bn of net investments was classed as Level 3, valued with inputs that are not observable in markets; this includes holdings in investment subsidiaries reported at that level as a whole (PDF p. 118). CPP Investments calls the resulting values "particularly judgmental" (PDF p. 117). For Level 3 private equities, 81% is valued with earnings multiples, discounted cash flows and comparable pricing, and 19% with net asset values supplied by fund managers (PDF p. 77).
Its own stress tests do not rely on smooth marks. CPP Investments estimates that a repeat of the April 2008 to March 2009 market moves, applied instantly, would cost about 15% or $119bn, and a repeat of the 2022 rate-hike cycle about 10% or $78bn, as at 31 March 2026 (PDF p. 49). The 2008 scenario includes the currency gains that typically come when the Canadian dollar falls in a crisis.
What does the Canadian model mean for crash risk and compounded growth?
The design aims for equity-like long-term growth with less damage in crashes, using diversification, private assets and leverage together; its risk is in the parts a price chart does not show every day. Over the twenty fiscal years to March 2026, the published returns compound to 4.49 times the starting value, about 7.8% a year, against an arithmetic average of 8.1%, by our calculation across the two published bases. The 2009 loss of 18.8% needed about two good years to recover on fiscal-year-end values.
Three features change the crash picture compared with a listed 85/15 portfolio. Private holdings are marked to model and can fall later than listed prices. Borrowing magnifies exposure; CPP Investments defines leverage risk as "the risk that excessive financial obligations heighten market and liquidity risks during periods of stress" (PDF p. 138).
And the Fund is not selling to pay pensions: in fiscal 2026 it received $76.6bn from the Canada Pension Plan and transferred $54.6bn back (PDF p. 37), and the Chief Actuary's December 2025 review found both accounts sustainable at legislated contribution rates.
For comparison, GPIF holds four listed asset classes in roughly equal parts, 80% through index funds, and accepts that its quarterly losses show in full. Neither design is free of crash risk; they differ in where the risk sits and how quickly it is reported.
How does copio review this portfolio?
Copio's free portfolio checker gave a euro adaptation of CPP's 85/15 Market Risk Target an overall B on 3 November 2026: A for fees, B for long-term assumptions, and C for historical efficiency, crises and diversification. This grades a euro adaptation of CPP's reference mix, not the actual Canadian-dollar reference or the Fund. The recorded result also depends on the checker's data and assumptions; our S&P 500 comparison found unexplained ETF-dependent differences in some readings.
The actual portfolio, with about 42% in private equities and real assets, 9.2% in credit and recourse leverage of 33.0% in the base account, cannot be entered cleanly: listed funds would have to stand in for private holdings and for the borrowing. The 85/15 target can, because CPP Investments defines it as two listed asset classes, global public equities and Canadian government bonds.
| Weight | ISIN | Fund |
|---|
| 85% | IE00B6R52259 | iShares MSCI ACWI UCITS ETF USD (Acc) |
| 15% | IE00B4WXJJ64 | iShares Core € Govt Bond UCITS ETF EUR (Dist) |
The euro adaptation is entered as a French securities account (compte-titres); both funds are registered for sale in France. MSCI ACWI covers developed and emerging markets, the broadest reading of global public equities. Euro government bonds stand in for Canadian government bonds, as the home-currency sovereign bonds of a euro investor.
Overall B; fees A 0.18% a year; long-term assumptions B −1.12 points a year; historical efficiency C −0.74 points a year; crises C −31.34%; diversification C, 2 sources of growth. Source: copio portfolio checker, engine portfolio-score-v11, measured 3 November 2026; figures translated from the French checker.
Fees: A, 0.18% a year. That is €18 a year per €10,000 invested, funds and contracts included. Two index funds with ongoing charges of 0.20% and 0.07% a year average 0.18% at these weights. CPP's operating expense ratio alone was 23.1 basis points in fiscal 2026, before investment-related and financing costs.
Long-term assumptions: B, −1.12 percentage points a year. The figure is the gap in expected annual growth to the optimal-growth portfolio. Under copio's capital market assumptions, the euro adaptation's expected growth is 6.7% a year against 7.82% for the model's optimal-growth portfolio, before inflation. With 85% in equities the mix already sits near the growth end; part of the gap is the choice of a 15% bond sleeve, and the figure is a model estimate, not a forecast.
Historical efficiency: C, −0.74 percentage points a year. The figure is the gap in annual return at equal risk over ten years, against MSCI World. Over the ten years to 2 November 2026, the euro adaptation's lines returned 10.46% a year in euros, against 11.19% for MSCI World scaled to the same risk with the rest in euro money market. The benchmark holds developed-market equities at the euro adaptation's risk, so the C means the euro adaptation's lines earned less for the same risk over that decade.
Crises: C, −31.34%. The figure is the estimated fall in a −40% shock to US equities. In that shock, the equity line accounts for −30.95 points of the −31.34% estimate and the bonds for −0.39 points. Euro government bonds cushion the fall by their weight, but in the model they do not rise to offset it. CPP's own estimate of about 15% for a 2008 repeat is a different scenario, applied to the actual Fund with its private assets, currency gains and leverage.
Diversification: C, 2 sources of growth. If world developed-market equities stop, expected growth above cash falls from 3.9% to 0.2% a year. The checker counts two sources, with most modelled growth assigned to equities. Its displayed "developed-market equities" label is incomplete for this input: ACWI includes emerging markets as well. Treat the stress result as a model convention, not evidence that the actual reference mix has only developed-market exposure.
What the euro adaptation leaves out is most of what makes CPP distinctive: private equities, real assets, credit, recourse leverage, active strategies, the Canadian-dollar base and CPP's selective currency hedging. You can enter the same euro adaptation, or your own portfolio, in the free portfolio checker.
What does copio think of this portfolio?
CPP shows why a risk target, a holdings list and a funding plan answer different questions. Its 85/15 target describes market risk; the implemented portfolio adds private assets, credit and financing. The fund compounded about 7.8% a year over twenty fiscal years, but lost 18.8% in fiscal 2009 and carries liquidity obligations from borrowing and pension transfers.
Fiscal 2026 contributions exceeded payouts by $22.0bn; that annual net inflow does not prove there were no asset sales or funding pressures within the year. The euro 85/15 example above helps examine a simple fund mix, not reproduce CPP's strategy or grade its managers.
Three checks for your portfolio
- Write down both net wealth and gross exposure. List investment assets and debts separately, then identify funds using borrowing, futures or swaps from their factsheets. For any leveraged holding, record the financing cost and what could trigger a margin call or forced reduction. CPP's 33% base-account recourse leverage is a reason to inspect these obligations, not a borrowing target.
- Match spending and commitments to liquid assets. List withdrawals, debt payments and any private-fund capital calls by date. Mark which holdings can fund them promptly and which have lockups or notice periods. Repeat the exercise after an equity fall and with private valuations unchanged: a stable reported value does not make an asset spendable.
- Choose a benchmark before judging complexity. Write a simple reference mix in your spending currency that reflects the risk you intend to take. Compare the actual portfolio with it over the same dates, after fees and financing, using both compounded growth and drawdowns. If private marks arrive later, compare longer periods as well as individual years before crediting apparent stability to diversification.
The portfolio checker can explore a listed-fund reference mix; our Yale study explains why the household implementation differs from an institution's.
