What fund reporting measures
- and what it doesn't

Can the best mathematician in the world calculate an IRR by hand? No. Can a fund with a 50% IRR return less money than one with a 15% IRR? Yes. Does the average wealth advisor understand how carried interest is actually calculated? Almost certainly not.

These aren't provocations - they're the starting points of a conversation between Emmanuel Delaveau (General Partner at Partech, former LP at CNP Assurances), Ludovic Phalippou (Professor of Financial Economics at Oxford), and Geoffroy De Cooman (individual LP and co-founder of Clariteer), recorded by Solenne Niedercorn-Desouches on the Finscale podcast.

45 minutes, three perspectives, zero concessions on the mechanics that most of the industry takes for granted.

The full episode is in French (45 min). English subtitles available on YouTube.

IRR: a fictitious rate that nobody can compute by hand

Ludovic set the tone early: "I can't explain how IRR is calculated, because nobody can calculate it by hand. The best mathematician in the world cannot compute an IRR manually." It's a fixed-point equation solved by iteration - by a computer, not a formula you can work through with pen and paper.

This matters because IRR is presented as a rate of return, and most people interpret it as one. If the fund shows 20%, the natural reading is "my money grew by 20% a year." That's not what it means. A fund with a 50% IRR can return less money than a fund with a 15% IRR. IRR rewards speed - getting cash back early - more than it rewards magnitude.

Ludovic offered a practical rule of thumb: IRRs between 0% and 15% are roughly trustworthy as approximations of actual wealth accumulation. Negative IRRs are never shown (because they produce nonsensical results). IRRs above 15% are also unreliable - but those are always shown.

Three practices that move the IRR

The episode identified three practices - each with legitimate operational reasons - that happen to have a significant side effect on IRR:

Subscription credit lines. The GP borrows from a bank to fund investments, delaying the actual capital call to LPs. When the first distribution comes back, the GP repays the bank and calls the capital almost simultaneously. The LP's money was committed for a much shorter apparent period, and IRR - which is hypersensitive to timing - jumps mechanically.

Early exit of top performers. Sell your best asset as fast as possible. This locks in a large positive cash flow early in the time series, which disproportionately inflates IRR. As Ludovic put it: "You keep your losers in the portfolio and exit your winners as fast as you can. Everyone does it."

Track record cherry-picking. When raising a new fund, a manager can attribute deals from a prior employer to their personal track record. An early positive cash flow from someone else's fund, grafted onto the start of your time series, produces an impressive headline IRR.

The structural US-Europe gap is not (only) about performance

US funds systematically show higher IRRs than European funds, which show higher IRRs than emerging market funds. But this hierarchy doesn't necessarily reflect a performance gap - it reflects unequal access to the tools that amplify IRR.

US banks are eager to provide subscription credit lines (secured against LP commitments, not the portfolio itself). Selling a successful company quickly is easier in the deepest, most liquid market in the world. European banks are more cautious with credit lines; exits take longer. In Africa or Latin America, a bank won't lend against LP commitments at all. The same underlying performance, dressed with different tools, produces different IRR headlines.

Emmanuel added nuance from the GP side: the US market is also purely contractual - terms vary wildly from fund to fund. Europe has historically converged more, with the "European waterfall" offering slightly more LP protection.

The catch-up clause: a widespread misunderstanding

Ludovic raised what he called a "simple question" that trips up even senior professionals. If a fund has an 8% hurdle rate and 20% carried interest, and it returns 10% - how is the carry calculated?

Most people answer: 20% of the difference between 10% and 8%, so 20% of 2%. But nearly all funds include a 100% catch-up clause, which means the GP takes carry from zero, not from the hurdle. The difference in fees is enormous. And this is not an edge case - it's the standard structure.

"I have this conversation with very senior people in the industry," Ludovic said. "The fraction of people who understand this is very low. Among retail investors, it's essentially zero."

The hidden assumption behind every IRR

This was the central insight of the episode.

Geoffroy walked through it step by step: if you buy an asset for 100 and sell it for 300 in 10 years, the annualized return is straightforward - gains divided by investment, raised to the power of 1/n. No debate.

But a fund has many cash flows. To get back to that simple ratio, you need to discount calls backward and compound distributions forward. That introduces three unknowns: the overall return, the rate on uncalled capital, and the reinvestment rate.

The industry's solution: assume all three rates are equal. One equation, one unknown, solvable by computer. Elegant - but it means a 25% IRR implicitly assumes your uncalled capital is compounding at 25%, and every distribution is immediately reinvested at 25%.

Geoffroy made it personal: "Gregory and I both invested in the same fund at Partech. But Gregory has a higher risk appetite - his uncalled capital sits in an ETF, earning about 8%. I'm more risk-averse, so mine sits in a term deposit at about 2%. When we each discount our calls at our own rate, we get very different numbers. IRR is personal - it depends on the investor's context."

For a deeper dive into this topic, including the formulas and a comparison table: Your IRR is not my IRR.

The expense ratio gap

In public markets, decades of standardization produced the Total Expense Ratio - a single number that lets investors compare fees across mutual funds. Private equity has no equivalent. Ludovic described participating in a project to create a reporting template: they reached 150 columns and still didn't have enough.

Even something as fundamental as knowing how much you're paying in fees is, in his words, "beyond what most people can simulate - even with the 200-page LPA in front of them." He estimated that only three investors he knows of have built the software to do it properly.

"Democratization" - a word worth questioning

Ludovic made a linguistic point that landed: the industry calls retail access to private equity "democratization." But opening a complex product to people who can't evaluate it has nothing to do with democracy. "If we legalize cannabis in France tomorrow, we won't call it the democratization of cannabis."

The point wasn't to oppose retail access - he explicitly said people should be allowed to invest in what they want. The point was that the presentation of returns and fees needs to be held to the same standard as public markets if the audience is broadening. Today, the regulatory gap between mutual funds (heavily regulated on disclosure) and private equity (barely regulated on presentation) is enormous.

What ideal reporting could look like

The episode closed with each guest's vision:

Emmanuel emphasized DPI - actual cash returned - and the conversion rate: when unrealized assets are eventually sold, how close is the exit price to what was reported in the quarterly NAV? That measures the quality of the GP's marks, not just their stated performance.

Geoffroy described a vision where all private market positions sit on a single screen, up to date, net of fees, and comparable. He also pushed for GPs to share more forward-looking cash flow hypotheses: "Today every LP builds their own 10-year model in Excel, which is necessarily less informed than if it were based on the fund's own assumptions." The diversification effect across a portfolio means individual forecast errors partially cancel out - but only if the data exists. This is what he is building at Clariteer.

Ludovic is working on a new return index that he hopes will be more reliable than IRR. He acknowledged the difficulty: "If I just manage to make a proposal for measuring returns that's roughly reliable, that would already be something." He's been working on performance measurement problems for over 20 years - and has recently reactivated a project on the duration question specifically.

Guest recommendations

• Emmanuel: Pitch Anything by Oren Klaff - on the psychology of storytelling and narrative in a financial context
• Geoffroy: Measuring Private Equity Fund Performance, INSEAD background note, 2019
• Ludovic: His podcast and upcoming book Valuation in Private Markets (details at pelaidbare.com)