CalPERS board ponders the risks of TPA

CalPERS CIO Stephen Gilmore talks the board through their role in setting the risk parameters behind a Total Portfolio Approach. The investment team hope the board will have selected its level of risk tolerance by November off which a TPA strategy can launch in July 2026.

Stephen Gilmore, chief investment officer of CalPERS, used a cooking analogy to describe the differences between the Total Portfolio Approach (TPA) that he hopes the pension fund’s board will introduce to manage the $533.4 billion portfolio and the strategic asset allocation (SAA) it currently relies on.

Speaking during the February board meeting, Gilmore explained how a SAA involves checking in every now and again (CalPERS adopts a SAA to determine its investment strategy every four years) and optimises at individual asset class levels. It is like cooking with a recipe book that relies on specific ingredients, he said.

In contrast, TPA is more continuous and optimises at the whole portfolio level for more efficiency. It is more innovative; the team are tasked with an objective and can use their discretion to provide “the same nutritious food.” Gilmore joined CalPERS in July last year from NZ$76.6 billion ($46.7 billion) New Zealand Superannuation Fund where he oversaw a TPA approach.

Under TPA, CalPERS’ board would set broad goals for managing the entire fund and give staff the task of implementing the strategy with one reference portfolio-type benchmark (under the current SAA, CalPERS has 11 different benchmarks) as well as a risk budget. Investments would be made according to whether they contribute to the desired outcome of the total fund rather than if they help fill out the asset class target allocation.

Measuring the Risk Appetite

An essential building block of the strategy involves ascertaining the board’s risk appetite. By this November, the investment team hope the board will have selected its level of risk tolerance off which it will be able to construct a portfolio to go live in July 2026.

CalPERS board would set its risk appetite by selecting a passive reference portfolio of stocks and bonds with active risk limits. The reference portfolio excludes alternative asset classes and alpha strategies such as private equity and private real estate.

The reference portfolio expresses risk tolerance and does not include the actual or target portfolio positions. The objective of an actual portfolio is to outperform the reference portfolio by using expertise, including additional asset classes and alpha-generating strategies. The actual portfolio will have illiquid assets, and won’t be mark to market so the observed tail risk will be lower.

In previous meetings, CalPERS’ board expressed its desire to increase risk to reap the benefits of an improved funded ratio and lower contribution rates – but not be exposed to large drawdowns. Board members agree they are focused on a total return objective rather than peer comparisons or relative returns. They are also open to innovation, more internal management, additional complexity and the higher costs that come with that.

Gilmore flagged the trade-off between shooting for additional returns and exposure to downside risk. “Tail risk increases as you increase the equity exposure,” he explained. The Board has expressed its concern that drawdowns and declines are not too long.

He also warned about the risk of macro- economic scenarios, particularly stagflation, on a future portfolio with the same level of risk as a typical 70:30 portfolio. Stagflation would impinge on equities (low growth) and bonds (high inflation) creating a potentially damaging scenario, he warned.

CalPERS board also need to explore their risk appetite regarding any deviation in portfolio returns from the discount rate. Gilmore flagged that with a 70:30 portfolio, returns can deviate from the discount rate over sometimes lengthy periods “It’s worth noting that 5-7- and 10 year rolling windows had incidents of negative returns,” stated his presentation.

A single benchmark

A new reference portfolio would be low cost and “simple” comprising easy-to-explain sources of risk and return and built using bonds and equities – the two most scalable and liquid asset classes.

Under its current SAA, CalPERS currently has 11 different benchmarks. Gilmore reflected that it is sometimes hard to see if the team have done a good job with so many benchmarks because they create different nuances. “With a reference portfolio it is much simpler; the question is: ‘Has management done better than a simple liquid portfolio,'” he said

Gilmore listed a hierarchy of risk in a TPA. The board’s overall risk appetite – and how much market risk they want – sits at the base of the pyramid. Next comes asset classes and their relevant weights, followed by manager selection. He explained that assets would be carefully matched to the level of risk the board is comfortable with.

Education of stakeholders outside the board meetings will continue with webinars.