Why suffer at the hands of the denominator effect when the true market value of private assets is available?

When one asset class in a portfolio underperforms (or outperforms) markedly, the resulting change in the total portfolio value (the denominator) can make allocations to other asset classes significantly overweight (or underweight). Indeed, strong reported performance in 2021 pushed several asset owners over their limit for private assets and private equity (PE) in particular. The bear market of 2022 in stocks and bonds markedly added to this effect among institutional investors. For instance, the Kansas Public Employee pension plan has reached 11% alternatives, mostly PE, when its target was 9%.

Other examples include the Oregon Public Employee Pension Plan, which saw its PE allocation reach 27% in 2022. Likewise, the State of Wisconsin Investment Board is running up against its 15% private asset allocation limit, having seen a 47.5% rise in its PE assets under management (AUM) in 2021. In May 2022, the Ohio Bureau of Workers’ Compensation reached a statutory cap and required a waiver to rebalance its real assets.

In Europe, this phenomenon has taken on a more systemic form as part of the liability-driven investment (LDI) crisis whereby, to cope with the liquidity crisis, many pension funds have quickly liquidated their listed equity and bond assets and find themselves with private asset values that can exceed 25%. This is the so-called denominator effect. This normally calls for a rebalancing of the portfolio – buying and selling securities to achieve the desired target allocation

However, when more liquid public markets fall and allocations to private, relatively illiquid assets become too large, rebalancing the portfolio can be difficult. Building or unwinding private-market positions takes time and costs are high. Investors might be tempted to ‘wait and see’, but the consequences of doing nothing are not benign. First, strategic asset allocations are rendered meaningless since investors do not know if and when over or under-allocations will change again.

Second, investors in private markets need to slow or stop committing capital to asset managers, which complicates private investment programmes both for general partners (GPs) and limited partners (LPs).

As a result of the denominator effect, many pension plans are pacing their commitments to private asset managers, preferring to stick with existing ones, even stopping investing if statutory caps on alternatives are reached. LPs also report wanting to sell stakes in the secondary market while valuations hold. As a result, pension funds are now the largest sellers of PE stakes, ahead of GPs facing the prospect of suffering large net asset value (NAV) discounts. Over Q3 2022 it was not rare to see haircuts on private asset of 10-15%. On the secondary private-equity buyout market, discounts that were below 5% in 2021 can now surpass 10%.

Asset managers are now seeing lower commitments in an environment in which expectations for investments in new funds are also being dampened by the macro-cycle. The 2015-18 vintages were good years, but the managers of newer vintages may have to hold on to assets longer in a higher-interest-rate environment that will depress exit values. This last point begs the question: if investors expect lower private exit valuations tomorrow because discount rates have increased, why not re-value private assets to reflect their current market value today and avoid a large part of the denominator effect?

Take for example, the infra300 index, which tracks the performance of 300 unlisted infrastructure companies worldwide on a mark-to-market basis, that is, reflecting monthly the impact of the yield curve and a risk premium calibrated to reflect the latest investor preferences for unlisted infrastructure assets. Unsurprisingly, with an asset class that has a duration of 10-15%, higher interest rates and higher risk premia have had a depressing effect on market prices; the infra300 was down 5% on the year as of Q3 2022 (local currency returns).

While this is not as bad as a comparable index of global equities (about -15%) or bonds (about -20%), such lower market valuations for private infrastructure would go a long way to minimising the denominator effect.

The lag and staleness of private markets is well known. For all their adherence to high-level ‘fair value’ principles, appraisals do not reflect the current market value of private assets, causing a large part of the denominator effect.

This should not be acceptable to either LPs or GPs, who suffer from such a discrepancy that could be resolved using market data and adequate asset pricing techniques. Private asset valuation methods that have been designed for the purpose of financial reporting are, for at least three reasons, only now having significant negative consequences in the area of risk management.

First, as private assets take up more space in the portfolio, it is necessary to pay more attention to managing these risks. Until now, the marginal role of private assets allowed their complete decorrelation with capital markets to be assumed, even though this was more the result of the lack of serious data than any economic or financial reality (ie, risk managers could not calculate correlation matrices in universes including private assets). Today, it is necessary to focus on the diversification potential of private assets among themselves and with the broader portfolio. Hence the need to model their risks and test this modelling with good data.

Second, if one carries out rebalancing to manage risk contributions or to maximum exposures by value of asset class, it is necessary to know the value of these assets to measure their risks. Since private appraisals have large backward-looking biases, unlike listed assets, they do not immediately reflect these risks and/or maximum exposures. Moreover, their low frequency does not allow ‘betas’ or individual risk exposures to be measured, such as interest-rate risks, which are absent in appraisals that rely on smoothed historical rates. With ‘proper’ fair-value marks using the latest interest rate and risk premia estimates – risk management would be possible, as well as a fair assessment of the denominator effect.

In addition, proper estimation of private asset (notably infrastructure, but also private debt) sensitivity to interest-rate risk could lead to its status being reconsidered in asset-liability allocation. In effect, derivative-based LDI strategies allowed the investor to avoid the question of the liability nature of their assets, since the derivatives ensured perfect matching. Today, the desire to deleverage pension fund balance sheets has led to the return of hedging interest-rate risk with cash assets. There is a risk that this ‘cashisation’ of liability hedging will lead to a mechanical reduction in the share allocated to private assets if one does not measure their interest-rate risk exposure seriously, giving rise to an allocation with a considerable majority of bonds.

Third, any optimal rebalancing strategy has a trade-off between the benefit of the change in risk parameters and the cost of this rebalancing – to do this it is necessary to estimate the costs of rebalancing. Without an assessment of the values of transactions on the secondary market at that time, this cannot be done. Knowing the fair value of private assets is an important subject in implementing an effective, dynamic risk-management strategy.

Today, thanks to advances in private-asset data and valuation technology, risk managers and asset managers do not have to suffer from the denominator effect, unless they prefer to wait it out or sell assets at steep discounts, in which case we propose to call it the denominatrix effect.

by Noël Amenc, Associate Professor, EDHEC Business School, Director, Scientific Infra

Frédéric Blanc-Brude, Director, EDHEC Infrastructure Institute, CEO, Scientific Infra