We develop an expected credit loss (ECL) framework that enables institutional PE investors to quantify embedded credit risk during the J-curve trough, when recommitment decisions are made and traditional metrics are least informative. The framework complements rather than replaces traditional metrics, adapting ECL methodology from banking regulations to provide fund-level comparison on a credit-adjusted basis. Illustrative calibrations suggest embedded credit risk varies meaningfully with vintage year, sector concentration, market conditions, asset composition, and general partner (GP) quality, with peak-vintage funds embedding 500–800 basis points of risk invisible to traditional IRR. We develop actionable thresholds for portfolio construction, performance evaluation, and GP selection, though external validation requires proprietary portfolio company default data that is not publicly available.
Charoenwong et al. (Tue,) studied this question.