CPP: Why doing the right thing in TPA can look like losing

CPP Investments has argued that a TPA portfolio has a high chance of appearing to have underperformed if evaluated against traditional SAA benchmarks, even when the underlying investment decisions are sound.

In a recent paper, the C$793 billion ($564 billion pension giant made the case for a “multidimensional” performance framework to assess TPA portfolios that consider factors beyond absolute financial returns, such as risk-setting, diversification and resilience.

The paper argued that the composition of TPA portfolios “intentionally differs” from traditional benchmarks in pursuit of diversification, resilience and dynamic positioning, a divergence CPP said can produce substantially higher tracking error against a single benchmark.

“Because TPA portfolios intentionally depart from conventional benchmarks, short-term benchmark-relative underperformance remains plausible even when portfolio decisions are sound and expected long-term outperformance is intact,” the paper said.

In a stylised demonstration of the dynamic, CPP Investments said a TPA portfolio can have higher alpha expectations compared to SAA but larger error bands against the benchmark due to substantially lower correlation of its returns.

Both SAA and TPA portfolios have error bands with a chance of “false negatives”, where the portfolio underperformed the benchmark but has positive true alpha.  

“Over a single year, both the SAA and TPA portfolios show significant incidence of underperformance to the benchmark,” the study said, but added that over five and 10-year horizons SAA portfolios are increasingly unlikely to produce false negatives while the chances remain elevated in TPA portfolios.

“[It] is not intended to suggest that either the SAA portfolio or the TPA portfolio is inherently superior to the other. Rather, it illustrates a measurement challenge that arises when portfolios are intentionally diversified away from a single benchmark.”

Apart from the false negatives, another problem is that benchmarks can shift significantly over time due to movements in company market cap and hence concentration, creating a risk that TPA portfolios are trying hard to diversify from.

“This creates a distorted incentive problem. Benchmark-relative evaluation can reward concentration when concentration is winning and penalise diversification that is being maintained to support long-term resilience,” the paper said.

“Over time, this can pull portfolios closer to the benchmark – even when doing so undermines the broader objective of diversification and total portfolio resilience.”

The paper came after CPP Investments said it underperformed its benchmark in the fiscal year of 2026, citing similar reasons. In the year to March 31, the fund’s benchmark portfolio returned 13.2 per cent while CPP Investments delivered a 7.8 per cent net return.

Significant concentration of large-cap technology stocks in public equities has driven the benchmark performance, while CPP’s “more diversified asset mix across public and private markets, sectors and geographies that… limited participation in strong equity market rallies,” the fund said.

CPP Investments’ paper further argues the fundamental framework for evaluating portfolios built under the two different approaches should be different. “In a sense, SAA is like a prize fight pitting the chosen portfolio design against a single competitor, the benchmark. Management can study this lone competitor at length, hugging or replicating the benchmark where they possess no edge and separating and fighting where they sense advantage,” the paper said.

“TPA places the chosen portfolio design in combat with all comers, aiming to outperform all feasible alternatives in the service of institutional aims.”

The fund proposed that TPA portfolio be evaluated against total return outcomes, risk and capital allocation, portfolio construction and diversification, investment selection, decision quality and process, and regime resilience.

This multi-focal framework can lead to different judgments of investment decisions under TPA. For example, a single strategy may outperform the benchmark in financial returns but increase concentration risks or add duplicate exposures to elsewhere in the portfolio.

“Benchmarks remain essential tools for attribution, discipline and accountability, but under TPA they become diagnostic inputs rather than singular verdicts on success or failure,” the paper says.

“The question is no longer simply whether a portfolio outperformed a benchmark, but whether all management decisions, including portfolio design and execution, improved the delivery of long-term institutional goals.”