The Receipt
CPP Investments, the manager of the Canada Pension Plan's fund, holds roughly 12% of its assets in Canada. Its chief executive, John Graham, reportedly attributed the low share to an "opportunity deficit," saying the fund would deploy more capital at home, as he told the Financial Times, "if we see opportunities." The claim shifts the question from CPP's allocation choices to Canada's supply of suitable investments. It is worth testing that broader diagnosis against the investment record.
The record supports a broader version of the diagnosis, though not every use the debate makes of it. Real business investment per worker in Canada has fallen since the mid-2010s while it climbed in the United States. By one construction, Canadian workers now receive roughly half the new capital of their American counterparts. Statistics Canada attributes weaker productivity growth since 2015 "largely" to weak capital investment, and its firm-level work shows the decline was concentrated in large and medium-sized firms, not small ones. The shortfall is sharpest in the categories most tied to future output: machinery, equipment, and intellectual property.
This is not a finding that CPP is right to hold only 12% at home. The fund's legal mandate is to maximize returns without undue risk across its whole portfolio, not to develop the domestic economy, and a fixed domestic-share target is the wrong yardstick for that mandate. Nor do the aggregate figures support a claim that CPP has steadily withdrawn capital from Canada: its Canadian holdings are worth more today than a decade ago. The finding is narrower, and it turns on a distinction the debate tends to collapse. What the data measures is a capital-formation deficit, meaning how much new capital Canadian businesses put in place per worker. What Graham was reportedly describing is an investability deficit, meaning whether enough Canadian assets meet a global pension fund's tests of scale, liquidity, governance and risk-adjusted return. The first is documented, and it makes the second plausible. It does not verify it. Nothing here establishes that CPP lacked additional Canadian investments meeting its mandate, and nothing here argues the fund should hold a particular share of its assets at home.
In January 2026, Ottawa was pressing Canada's pension funds to keep more money at home, chasing upwards of $500 billion in new financing for infrastructure and industrial projects as growth softened and exports fell, according to Financial Times reporting relayed by Benefits and Pensions Monitor. [1] The pressure had a specific target. CPP Investments manages the national plan's fund on behalf of roughly 21 million contributors and beneficiaries, and it holds about 12% of its assets in Canada, a figure the fund itself cites in its public materials. [2] Nearly half its portfolio sits in United States assets. [1]
Asked why, the fund's chief executive gave an answer that has become shorthand for the whole debate. John Graham told the Financial Times he was "super encouraged" by federal efforts to create large-scale investable projects but pointed to an "opportunity deficit," saying the fund would deploy more capital in Canada "if we see opportunities" while stressing its mandate to maximize return without undue risk of loss (Graham's remarks to the Financial Times, as reported by Benefits and Pensions Monitor). [1] The phrase reframes the question. It moves a portfolio decision into a statement about the country: the fund would allocate more domestically if it found additional Canadian opportunities meeting its scale and risk-adjusted-return requirements.
Part of that claim can be checked, and part cannot. Whether CPP specifically could have found more large, risk-appropriate deals is a counterfactual about one fund's deal flow, and no public dataset settles it. The wider environment is testable. If Canadian business investment conditions have weakened, that should show up in what businesses across the economy actually put in place, not only in what one pension fund chooses to hold. Those are different questions, and the rest of this piece answers the second one.
The share fell. The dollars did not.
The first thing the record clarifies is what "12%" does and does not mean. CPP's Canadian weighting has dropped sharply as a share of the fund: from 19.1% at the end of its 2016 fiscal year to 15.7% in 2021 and 12.0% in both 2024 and 2025, according to the fund's annual disclosures. [3] The dollar value followed a different path. It rose from $53.3 billion in 2016 to $78.3 billion in 2021, fell to $73.9 billion at the March 2024 fiscal year-end, and stood at $98.3 billion in CPP's portfolio disclosure for December 31, 2025, a calendar date rather than a fiscal year-end. [3] The percentage fell steadily. The dollar value was substantially higher at the end of the period than at the start, despite the decline from its 2021 level.
This forecloses one common framing. The aggregate figures do not support the claim that CPP has steadily withdrawn capital from Canada, whatever the falling percentage suggests. The percentage fell largely because the denominator grew faster than the numerator: total assets expanded with contributions, foreign returns and currency movements, while the Canadian book grew more slowly. A precise decomposition of how much each factor contributed is not available from the public disclosures, so the share alone cannot carry an argument about intent or withdrawal.
It also complicates the mandate question in a way worth stating plainly. CPP's governing statute directs it to invest with a view to a maximum rate of return without undue risk of loss, judged across the entire portfolio. [4] A domestic-content quota is not part of that mandate, and the fund's own rationale leans on scale: Canada is about 3% of the global economy, its materials argue, so a globally diversified fund cannot concentrate at home. [2] Notably, CPP's public explanations move between two different denominators. It cites 3% of the global economy in one place and 3% of world capital markets in another, and the two are not the same measure. [2] That slippage is a signal worth holding onto: the debate over CPP's domestic weight is fought with inconsistent yardsticks, and the "opportunity deficit" is one more yardstick that needs defining before it can be weighed.
One scope note belongs here. CPP is not representative of every major Canadian pension fund's domestic weighting, and several peers hold substantially larger Canadian shares, according to reporting on the sector. [1] It is the focus of this piece because its chief executive supplied the specific claim being tested and because it manages the national plan, not because its geographic allocation stands in for the sector.
The weakness Graham describes is measurable
Strip away the pension politics and the underlying claim is about Canadian investment conditions. On that, the primary record is consistent and unflattering.
Statistics Canada has stated that slower labour-productivity growth since 2015 was "largely due to weak capital investment," and that investment per worker in 2022 sat nearly 20% below its 2014 level. [5] By early 2024, real non-residential structures were about 25% below their peak and machinery and equipment about 17% below. [5] This is the national statistical agency describing the same environment CPP's chief executive describes, in the language of the capital stock rather than the portfolio.
The international comparison sharpens it. The C.D. Howe Institute, using OECD and Statistics Canada data, tracks business investment per available worker against peer economies. In 2014, Canadian workers received about 89 cents of new investment for every dollar their counterparts across the OECD received. By 2025, on the institute's estimates, that figure falls to roughly 70 cents, and against American workers specifically to 55 cents. Real investment per available worker dropped to about $15,000 by the third quarter of 2025 from a $19,400 peak in 2014, a decline of nearly a quarter. [7] Measured by capital stock rather than new investment, the average member of Canada's labour force had 9% less capital to work with in late 2025 than in 2015, with machinery and equipment down a further 20%. [7] A separate Fraser Institute analysis, using a simpler Canada-U.S. frame, reports real business investment per worker falling from $20,310 in 2014 to $16,493 in 2024 while the U.S. figure rose from $23,263 to $30,555, a slide from 87 cents on the American dollar to 54. [8] Using 2014, the pre-decline peak relevant to this article's decade window, makes the drop look steeper than the full series does: the same analysis puts the 2007 Canadian figure at $17,345, much closer to the 2024 level. [8]
These are constructed comparison frameworks rather than raw official series, and they use different denominators, currencies and start years, so they should not be spliced into a single trend line. What matters is that Statistics Canada's domestic series and two separately built Canada-U.S. and OECD comparisons point the same direction, [13] although those comparisons draw on common official data and are not independent replications. The direction is down, and it is down relative to the countries CPP could invest in instead.
One qualifier the honest version of this story cannot skip: the resource unwind explains part of the decline but does not exhaust it. The post-2014 collapse in energy-sector capital spending is real and pulled down the national figures, with Alberta, Saskatchewan, and Newfoundland and Labrador all recording steep per-worker declines over the decade. The firmest test comes from Statistics Canada's own firm-level work, which finds the pattern of decline among large, medium-sized and foreign-controlled firms persists after excluding mining and oil and gas extraction, and which describes post-2015 investment weakness as pervasive across industries. [9] [5] C.D. Howe's older figures corroborate it: excluding oil, investment per worker in 2016 was 57 cents on the U.S. dollar, down from 66 cents in 2010, and Ontario and Quebec ran well below both the OECD and U.S. averages. [6] That falsifies a purely resource-driven account. It does not establish that resources were a minor contributor.
Where the deficit actually lives
The most useful thing the data does is locate the shortfall. Not all investment is equal for an economy's future: spending on structures builds capacity, but spending on machinery, equipment, and intellectual property is what makes existing workers more productive. Those are the categories where Canada has fallen furthest behind.
C.D. Howe's category breakdown for 2024 puts Canadian investment per worker at roughly 1.05 times the U.S. level for structures, essentially parity, but only about 0.41 for machinery and equipment and 0.32 for intellectual property products. [7] Statistics Canada's national accounts show the same pattern in the flow: in 2024, non-residential structures, machinery and equipment, and intellectual property products all registered small annual declines in real terms. [5] Canada is roughly keeping pace with the United States on buildings and falling well behind on the equipment and ideas that raise output per hour.
The intellectual-property ratio carries a measurement caveat that the machinery comparison does not. C.D. Howe notes that IP accounting is not identical in the two countries, so the 0.32 figure is less clean than the 0.41 for machinery and equipment. The institute tests narrower categories it considers more comparable and still finds large gaps: U.S. software investment per worker at roughly twice Canada's, and research and development at roughly four times. [7] The limitation is real, and the finding survives it.
The firm-level evidence rebuts a natural objection, the idea that this is a small-business problem or a story about firms too minor for a fund like CPP to notice. Statistics Canada found that investment per worker fell 20% from 2006 to 2021, and that the decline was concentrated in larger enterprises. Large and medium-sized firms accounted for about 90% of the total decline, well above their 65% share of investment in 2021. Foreign-controlled firms accounted for roughly 30% of the decline despite representing about 20% of investment. [9] The forces associated with weaker investment are visible among large firms, not only smaller ones, which is the size class whose assets would be scaled for a pension fund.
The rising total that settles nothing
Against all this sits a fact that complicates the picture: headline capital spending in Canada has been rising. Statistics Canada's February 2026 release puts non-residential capital expenditure intentions for 2026 at $401.2 billion, about 3.7% above the preliminary 2025 estimate in the same release. [10] An earlier release had put 2025 intentions at $388.6 billion, but intentions and later preliminary estimates come from different survey vintages and should not be read as a single series. On its face, a rising total looks like a country investing more, not less.
The headline figures measure something different from the per-worker series, and the difference is the whole point. The capital-expenditure totals are in current dollars, so a meaningful part of each year's "growth" is price inflation rather than real capacity. They combine businesses, governments, and institutions, so public and project spending can carry the total even when private business investment is soft. And they are national totals, spread across a labour force and working-age population that expanded rapidly over the same years. [11] More dollars divided among more workers can still mean less capital per worker. That denominator effect is not merely a statistical artifact: rapid workforce growth depresses capital per worker unless investment keeps pace, which is part of the mechanism being measured rather than a distortion of it. C.D. Howe adjusts its available-worker count for the recent undercount of temporary residents and reports that the declining trend holds, including when employment is used in place of the labour force. [7]
The composition of the increase matters more than its size. Statistics Canada's release anticipates most of the asset-type growth in structures while machinery and equipment slips about 0.6%, with transportation, utilities, public administration and resource extraction contributing to the total and manufacturing remaining weak after a 2.6% decline in 2025. [10] The release also expects private-sector investment to recover slightly in 2026 after weakening the year before, so this is not simply a public-led increase. What the topline does not establish is a recovery in real business investment per worker, particularly in the productivity-enhancing categories. A nominal survey covering public and private organizations, growing fastest in structures while equipment spending falls, is not evidence against the per-worker trend.
The strongest case that the deficit is overstated
A serious version of the counter-argument does not deny the numbers; it disputes what they prove. Three lines are worth stating at full strength.
The strongest is that national capital formation is not a measure of CPP's opportunity set. Gross fixed-capital investment counts new capital put in place by businesses. A pension fund does not buy new capital formation; it buys assets, including existing public companies, credit, real estate, infrastructure and private equity, many of which are unaffected by whether a manufacturer bought a new machine last year. A country can record weak equipment investment while still offering large, well-governed, liquid assets that suit a global fund, and it can record heavy capital spending by governments or closely held firms that produces nothing a pension fund could reasonably purchase. The per-worker data therefore describes the environment in which Graham's account arises. It does not measure the number, price, scale or governance of the Canadian assets available to CPP, which is the thing his claim was about.
The second is that a domestic-share target is the wrong benchmark, and that framing CPP's 12% as evidence of anything is a category error. Sebastien Betermier, a finance professor at McGill and executive director of the International Centre for Pension Management, has argued against mandating or directing funds' allocations, warning that blending national-development goals into a retirement-security mandate risks the independence that made the Canadian model effective (Betermier, in remarks to Benefits and Pensions Monitor). [12] He and colleagues developed the concept of an "investable window," the set of conditions under which a domestic project actually fits a pension fund's obligations: competitive risk-adjusted returns, sufficient liquidity, appropriate scale, strong governance, and low development risk. [12] On this view the "opportunity deficit" is not vague at all. It is the gap between the projects Canada offers and that window, and the fix is for governments to widen the window by de-risking projects and clearing regulatory delay, not to move the fund's allocation by decree.
The third is that the conditions themselves are changing, though the measures involved are at different stages. According to Financial Times reporting, the federal government removed the 30% cap on pension funds' investment in Canadian entities, established a Major Projects Office to coordinate and accelerate national infrastructure, and began considering changes to ownership limits on municipal utilities. [1] One barrier has been removed, one coordinating body created, and one limit is under review rather than changed. If those conditions continue to shift and CPP's Canadian holdings continue to grow, the fund's stated position, to deploy more here as opportunities meeting its tests appear, is being tested in real time rather than remaining an untested explanation. Betermier makes a sharper version of the same point: the "silent loss" from failing to attract foreign capital into Canada, he argues, can be an order of magnitude larger than any gain from forcing more domestic deployment. [12] On that framing the entire debate is aimed at the wrong quantity.
These arguments constrain what our finding can claim. They do not rebut it. That Canada's investment weakness is real and productivity-relevant is a statement about the national accounts; that a domestic-share quota is the wrong response is a statement about pension mandates. Both can hold at once. The data establishes the first. It takes no position on the second.
What the data settles, and what it does not
The cyclical objection deserves its strongest form before the finding is stated. Canada's 2014 peak was an unusually capital-intensive moment in a capital-intensive resource cycle, and measuring from a peak exaggerates the structural share of any subsequent decline. The later years absorbed a pandemic, supply disruption and rapid monetary tightening. Some of the shift may reflect an economy moving away from extraction toward activity that requires less fixed capital per unit of output. Statistics Canada has also noted that firms have moved spending toward intangible assets that the national accounts capitalize incompletely, meaning some real investment is happening where the standard measures do not fully see it. [5]
What survives that objection is narrower than the headline decline but harder to break. The weakness persists in the firm-level data after mining and oil and gas are excluded. Statistics Canada describes it as pervasive across industries. Machinery and equipment remain clearly depressed relative to the United States, and the software and research categories that C.D. Howe considers most comparable still show gaps of roughly two and four times. [9] [5] [7]
So the record establishes a capital-formation deficit: broad-based, persistent, not confined to the resource sector, and worst in the categories that raise output per hour. It does not establish the investability deficit that Graham was describing. Those remain two different claims, and only one of them is measured. The honest statement is that Canada's documented investment weakness makes his account plausible without verifying it, and that nothing in the national accounts tells us whether CPP passed on Canadian assets that met its mandate.
One further gap is worth naming precisely, because it is easy to overstate. Among the mandates and sources reviewed for this article, none identifies a single institution responsible for reconciling nominal capital-spending growth with declining real business investment per worker, or for reporting against an investment-intensity target. That is a statement about what this review found, not a finding that no such responsibility exists anywhere in government. Establishing that would require a survey of federal and provincial economic mandates that this piece has not done. Statistics Canada measures the decline. C.D. Howe and others benchmark it. The federal government has begun adjusting the conditions around it. Whether anyone answers for the aggregate is a question the record reviewed here leaves open.
What Would Change This Assessment
The finding rests on measured investment intensity, not on any claim about CPP’s deal flow. These are the specific results that would undermine it.
- Real business investment per available worker recovering to or near its 2014 level on a sustained basis, rather than a partial rise from the trough. As of the third quarter of 2025 it stood at roughly $15,000 against a $19,400 peak. [7]
- Evidence that the per-worker decline is an artifact of a fast-growing denominator rather than weak investment. C.D. Howe adjusts its worker count for the temporary-resident undercount and reports the trend holds using employment instead of the labour force; a credible measure showing otherwise would weaken the claim. [7]
- A finding that the weakness is confined to mining and oil and gas after all. Statistics Canada’s firm-level work finds the pattern persists when extraction is excluded; a reanalysis reversing that would narrow the conclusion to a resource story. [9]
- Revised national accounts showing the machinery, equipment and intellectual-property gaps closing once intangible assets are more completely capitalized. Statistics Canada notes current measures capture them imperfectly. [5]
- Disclosure from CPP quantifying the Canadian assets it reviewed and declined against its scale and return tests, which would move the investability question from unmeasured to measured.