US equity concentration poses portfolio dilemma for pension plans

Each month at BPM, we offer a slate of articles and content pieces that go deep on a particular topic. This month, we're focusing on Public Equities.

Canadian pension plan sponsors face a stark dilemma in their global equity allocations. While US tech earnings are the strongest in a generation, equity experts suggest the market concentration powering those returns is creating risks that are difficult to hedge and even harder to ignore.

“It’s not a stock market, it’s a market of stocks. The exceptionalism doesn’t come from people’s view on US politics or US capitalism. It comes from their size in the capital markets and the ability to use that size to create these really big companies,” said Nick Griffin, Australia-based chief investment officer at Munro Partners, part of CI Global Asset Management.

According to Griffin, network effects in the digital era funnelled users toward whichever platform was biggest, and the biggest platforms were American. That self-reinforcing cycle produced the Magnificent 7 and continues to favour US hyperscalers as the AI era takes shape.

“The US companies have ended up being the best in these areas, partly because of their size and partly because of their technical leadership,” he added.  “As we move into the AI era, that fragments a little bit from here. Still, the biggest hyperscalers are in the US. This concentration has always driven equity markets. Equity markets are always a function of just a few great companies and thousands of mediocre ones.”

While Griffin acknowledged exceptions, noting Taiwan Semiconductor Manufacturing Company (TSMC) leads in chip production and ASML dominates semiconductor equipment, he underscored that US firms hold the advantage through a combination of market size and technical leadership. He compared the current cycle to the mobile internet era that ran from 2008 to 2018.

“We still haven’t even got to like the really big apps of AI and what they’re gonna be,” he added. “It still feels like a five-to-10-year structural trend. And I think people are dramatically underestimating how big that could be for global earnings.”

Concentration risk hits governance hard for plan sponsors

As Sebastien Betermier, Montreal-based executive director at the International Centre for Pension Management noted, seven companies now account for roughly 35 per cent of the S&P 500’s total market cap and have driven more than 55 per cent of the index’s return over the past three years.

For anyone indexed to broad markets, he suggests that’s produced strong results but it also raises a pointed question about whether those same stocks are overvalued, and what it means to carry that risk in a portfolio with so few names doing the heavy lifting.

He cited a 2025 paper by University of Chicago researchers, discussed at the Northern Finance Association Conference in Quebec City, that made a counterintuitive case: the concentrated mega caps may be underpriced.

“Many funds are hitting their limits because there's regulatory limits,” Betermier explained. “If you're a mutual fund in the US, there's limits to how much you can own in select stocks. And if you take positions of stocks that are more than 5 per cent of the portfolio and you add all them up, it cannot exceed 50 per cent of the portfolio. But today, many of the funds are hitting that constraint. So you have funds that potentially would want to track the index and go more, but they can't.”

Betermier laid out the governance trap that concentration creates for pension fund managers. Diversification is supposed to be a foundational principle - one of the few uncontroversial ideas in finance. But when the benchmark itself is dominated by a handful of stocks that have outperformed everything else for three years running, the act of diversifying away from it generates tracking error. A fund that did everything right - moved into privates, tilted across asset classes, pursued mispriced opportunities - can still look like it failed simply because seven stocks carried the benchmark higher than anything else, he said.

Rethinking benchmarks and manager evaluation

As to how plan sponsors are responding to this level of concentration, Devin Forbes, associate director of investment consulting at Eckler in Toronto suggests many are staying the course by staying disciplined within their existing investment philosophy and rethinking how they evaluate managers.

“I don’t think it’s at a point where people are becoming extremely uncomfortable with the situation. I think there are enough opportunities out there, specifically in the global space, to construct a portfolio,” he said.

The global opportunity set, in his view, still offers enough breadth to build diversified portfolios that can outperform without taking on excessive risk. The difficulty is more acute within US-only mandates, but Eckler’s focus on global allocation provides a wider canvas, Forbes noted.

“We’re trying to focus on a global allocation here rather than just a single US outlook,” he said.

Earnings momentum makes the bull case

But Griffin sees the gap between earnings momentum and valuation as an opportunity rather than a warning sign, adding the mechanics supporting that view are twofold: AI-related capital expenditure is expected to continue, even if the pace moderates, funnelling more revenue to the companies building data centre infrastructure. Consequently, hyperscalers are sitting on a growing base of installed compute capacity that they have yet to fully monetize.

Looking ahead, he laid out a straightforward bull case for US equities heading into next year. Earnings estimates are tracking 25 per cent to 30 per cent higher this year, a pace he called unprecedented outside of a post-recession recovery. The market multiple has compressed in response, reflecting skepticism that the growth is sustainable.

“It’s very hard to paint a picture where earnings in the US don’t go higher again next year,” he said. “As the market starts to derate, that’s probably an opportunity [to revisit allocations].”

Aubrey Hearn, senior vice-president, portfolio manager and lead of US & small cap equities at Toronto-based CI Global Asset Management (GAM) also acknowledged that CI GAM is relatively bullish on US equities heading into next year. Valuations have come down from the 21 to 22 times P/E range at the start of the year to roughly 18 to 19 times, which sits in line with the 10-year historical average. That puts the market at fair value rather than bargain territory, but the setup is constructive, he noted.

However, the more interesting catalyst, Hearn suggests, is what happens as companies begin to deploy the AI infrastructure they have been spending on. He suggests there’s growing evidence of adoption, particularly as AI is appearing more frequently on corporate conference calls, and firms like OpenAI and Anthropic are showing revenue traction.

“As you see future evidence of AI savings in companies, more productivity enhancements, I think it could be a good year next year for US equities as that narrative kind of broadens out to sectors outside of tech,” Hearn said.

The CapEx dependency problem

But he cautioned against treating that as an all-clear signal. The distinction from 2000 cuts both ways: this is a CapEx-driven cycle, and the elevated market caps of many companies are directly tied to the continuation of AI-related spending.

If that spending were to slow or stall, the pain would not be confined to a few names - it would ripple through the broader indices given how much AI exposure is embedded in cap-weighted benchmarks.

“You don’t want to be the last person dancing either. You just have to be aware of what you’re buying when you’re buying these indices,” he added.

According to Griffin, the sheer global reach of companies like Microsoft and Amazon is what makes concentration so hard to escape. These firms are not just dominant in the US because they operate as default infrastructure for the world.

As a result, Griffin suggests that puts pension investors in a difficult position, noting the largest names command outsized index weightings at $4 to $5 trillion market caps, and even attempts to diversify into emerging markets end up tied to the same hyperscaler spending cycles.

For investors who tilt toward Taiwan or Korea, for instance, he suggests that still leaves a portfolio exposed to the CapEx decisions of the same US mega caps. Griffin acknowledged the discomfort this creates for plan sponsors but argued that concentration is not a flaw in equity markets, it’s how they have always functioned.

For pension clients who want to stay in public equities and remain close to the benchmark, he said, there is no realistic way around the concentration problem short of moving into entirely different asset classes.

Additionally, Griffin sees the concentration of returns in US mega caps as a structural advantage, not a vulnerability. He pointed to the natural cycle that reinforces his point, underscoring how scale attracts capital, capital funds innovation, and innovation draws more scale.

That cycle has held for more than three decades, going back to the start of digitalization, and he argued it would have been nearly impossible to replicate elsewhere. A platform like Google or Netflix needed the largest possible user base to generate the network effects that made it dominant, and the US provided that by default.

Policy noise versus regime change

While he acknowledged the current political environment makes matters difficult, Griffin doesn’t believe that federal-level policy risk defines the investment case for pulling away from the US, particularly as the competition between states creates a built-in release valve for capital.

“If California pushes one way, then Texas goes the other way,” he said. “Money can always find a home in the US where money's treated best. That can happen at the state level, not necessarily just at the US level, which also makes it a very capitalist environment and a good place to invest overall, because money likes to go where it's treated best.”

When it comes to pricing political risk, Griffin identified a specific threat most investors are ignoring:

“The biggest risk is definitely a pushback towards socialism, I think, in the 2028 election,” he said, arguing that tighter regulation and fewer data centre permits would slow the AI build-out.

Hearn took a different tack, suggesting the upcoming midterms could produce a split government that markets would welcome.

“Sometimes that stalemate can lead to more predictability. And the markets like more stability,” he said.