Existing stochastic discount factor methods for the valuation of private equity funds result in unrealistic time discounting. The authors propose and evaluate a modified method. Valuation has a risk neutral component plus a risk adjustment, and the authors fix the risk-neutral component by constraining the subjective term structure of interest rates with market data. They show that (i) the approach allows for economically meaningful measurement and comparison of risk across models, (ii) existing methods estimate implausible performance when time discounting is particularly degenerate, and (iii) this approach results in lower variation of performance across funds.