When a standard of measurement of returns allows close to 70% of investment managers (GPs) to claim their funds are first quartile performers (i.e. ranked in the top 25%) (1) – such as the case of the IRR – something is obviously wrong. Innate buy-side scepticism and academic evidences seem to have helped most investors filter and manage these marketing claims.

But there are other ways IRRs can more subtly fool private markets’ investors.

This is the case of the notion of the dispersion of returns (that supports the argument that only superior selection skills deliver adequate alpha).

The topic of the dispersion of IRRs is of particular relevance for advisers, gatekeepers and consultants. Again the financial report of Yale Endowment helps explaining the “fooling” point, based on Cambridge Associates‘ figures (because – again! – they do not mention they are talking of IRRs).

“Yale has never viewed the mean return for alternative assets as particularly compelling. The attraction of alternatives lies in the ability to generate top quartile or top decile returns. As long as individual managers exhibit substantial dispersion of returns and high-quality investment funds dramatically outperform their less skilled peers, Yale enjoys the opportunity to produce attractive returns for the Endowment and to demonstrate that manager alpha (excess return) is alive and well.”


Professor Phalippou in his paper “The hazards of using IRR to measure performance: The case of private equity” explains why the large dispersion of IRR is illusory. Performance appears more dispersed than it really is because of the reinvestment assumption for the interim cash flows that is implicit in the IRR definition.

“The re-investment assumption means that funds with high IRR have a higher IRR than effective rate of return and funds with low IRR (below re-investment rate) have a lower IRR than effective rate of return. It is thus mechanical: volatility is exaggerated.”

Comparing IRRs versus both listed markets’ CAGRs and median return produces only an illusion of eventually larger (but unrealistic) alphas….

At the cost of a potentially higher and painfully true fee….


(1) An Oliver Gottschalg's study is a much cited statistic, by the Economist and others (e.g. Tower Watson)