I am able to calculate the money-weighted return (XIRR equivalent in Excel) of my portfolio. Whilst I can compare this with ‘headline’ returns of ETF’s, Mutual Funds etc, I want to isolate the timing of my cashflows from my choice of investments. In other words, I want to see if I would have been better making my investment timing decisions and investing in the benchmark rather than investing in my own choice.

The difficulty seems to be that if I apply the same cashflows to benchmark the benchmark itself can be very misrepresentative – for example, if I have outperformed the benchmark, then make a cashflow deduction from my portfolio and try and replicate this in the benchmark (with the same dollar amount) then the benchmark return itself may go negative. Is there a common approach to this?


1 Answer 1


A similar problem arises in comparing private equity investments (which involve cash inflows and outflows) to the performance of a benchmark such as the S&P500.

The simplest method available, the Long Nickels PME, does what you say, it assumes that the same cash flows that occur in the PE investment are also made in the benchmark. The problem with this method is exactly what you mention: the amount theoretically invested in the benchmark can go negative if you withdraw a large amount of cash from your portfolio.

There are several solutions available in the literature, the one I would recommend is called the Kaplan Schoar PME. It is a ratio of discounted cash flows, such that if greater than 1 you are outperforming the benchmark, and if less than 1 you are underperforming. In the numerator you discount (using the rates of return of the benchmark) the cash outflows from your portfolio, and in the denominator you discount the cash inflows into your portfolio.

These methods are described in the Wikipedia article Public Market Equivalent PME link

A theoretical justification of the Kaplan Schoar PME method can be found in a recent article by Sorensen and Jagannathan SSRN link.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.