Factor-Based Portfolio Construction

Using a factor lens, institutions may target explicit exposures to risk factors¹ including growth (equity), interest rates, and credit, as a means of aligning the portfolio with long-term risk and return objectives. In this paper, we demonstrate how these targeted exposures can be quantified using both listed and private asset classes, with private assets proxied by privateMetrics® and infraMetrics® indices. A growth-oriented portfolio targeting 65/25/20 exposure to the equity, interest rate, and credit factors, respectively, is constructed across listed and private assets. We compare a simple portfolio comprised of listed assets with one that incorporates private assets, with both portfolios having the same allocations across the three risk factors.

Diversification within the Factor Allocation

Private equity, infrastructure equity, and infrastructure debt each exhibit return variation beyond common risk factors, and these unexplained components are uncorrelated with listed asset residuals. This offers potential diversification within the factor allocation and may contribute to improved portfolio Sharpe ratios while maintaining similar factor exposures.

Infrastructure as a Diversifier

Infrastructure equities offer diversification within a factor allocation framework. With statistically significant betas to both the equity factor (~0.36) and interest rates (~1.01), infrastructure equities contribute meaningfully to desired exposures, not easily replicated from equity or bond only assets. Similarly, infrastructure debt provides diversification relative to investment grade bonds, with lower equity beta and credit exposure than IG bonds, and differing rates beta. The uncorrelated residuals suggest that obtaining rates and credit exposure via infrastructure debt may enhance risk adjusted performance.

Time Varying Betas and Uncertainty

In our last TPA piece, we showed that the risk factor betas are time varying. There are periods when private equities and infrastructure equities may have higher/lower sensitivity to risk factors that are defined by tradeable listed assets. This creates a challenge for allocators seeking a certain level of exposure across equities, rates, credit, or other factors. Related to this is the uncertainty around the coefficients. For infrastructure equities, we find an average equity factor beta of ~0.36 using 20 years of monthly (~250 months of returns). At a 5% confidence level, the beta is 0.25 to 0.47. This can make it a challenge to understand true exposure. Extending this across other betas and unlisted asset classes and the confidence in the betas may become an issue. For institutions with large private asset allocation, estimation errors have real consequences. Using privateMetrics and infraMetrics helps reduce this uncertainty, as higher-frequency data that better reflects current private market conditions can provide more accurate estimates.

Footnotes:

¹For this paper, risk factors refer to Equity, Rates, and Credit. This distinguishes them from style factors.