The Covariance Structure between Liquid and Illiquid Assets

Abstract

The traditional methods of estimating investment risk based on the observed variation of asset prices falter on sparsely traded assets owing to the lack of price quotes. The price variance of illiquid assets such as private equity, loans, or real estate is, if calculable at all, not meaningful. Nor does the observed covariance between illiquid assets indicate a degree of proximity in the price behavior. The author proposes a cross-sectional approach instead, taking the price dispersion across assets as an estimate of variance and using the so-called cross dispersion of covariance. The covariance matrix computed in this way over traded and nontraded assets seems coherent with the (sub)matrix computed over time on traded assets. The cross-sectional matrix sets a universal framework for managing portfolios invested in traded and nontraded assets, making all-encompassing risk assessments possible.