Ludovic Phalippou is professor of financial economics at the University of Oxford’s Saïd Business School.
A decade ago, Steven Davidoff Solomon, a UC Berkeley corporate law professor, published in the New York Times an article entitled: “Private Equity Fees Are Sky-High, Yes, but Look at Those Returns”.
The message was simple. Retail investors were understandably eager to invest in private equity funds, given the incredible money they were generating:
As an asset class and with the right fund, private equity is nigh unbeatable, well worth the fees paid. (…) Over the last 20 years, Yale has had a 36.1 percent return from its private equity portfolio. (…) Super achievers in the top quartile include Kohlberg Kravis [sic], which has had a 26 percent net return for its long-term funds since inception in 1976, and Apollo, with a 25 percent net return since inception.
During such a long timeframe, these performance statistics are truly remarkable and obviously exciting to retail investors. In fact, the returns are so high that they surpass Warren Buffett’s record by a wide margin over any of these timeframes.
Even more remarkably, these return figures have remained just as strong a decade on. In its latest 10K annual filing, KKR said:
From our inception in 1976 through December 31, 2024, our Private Equity and Real Assets investment funds with at least 24 months of investment activity generated a cumulative gross IRR of 25.5%, compared to the 12.2% and 9.5% gross IRR achieved by the S&P 500 Index and MSCI World Index, respectively, over the same period, despite the cyclical and sometimes challenging environments in which we have operated [note: Solomon seems to have mixed up KKR’s net and gross IRR, as well as the firm’s name].
Similarly, Apollo brags in its 10K about the 39 per cent gross “internal rate of return” (or 24 per cent net) generated by its private equity funds from inception through to the end of 2024. Across almost all regulatory and marketing material filed by private equity, awe-inspiring IRRs are common.
Apparently, these private equity firms have managed to defy the laws of mathematics, economics — and reality.
Since most people typically don’t have a good grasp of compounding, it might be helpful to express these numbers in dollars to show how fantastical they really are. If KKR’s first $31mn fund from 1976 had compounded at 26 per cent a year it would be worth $2.6tn today. Add in its second $350mn fund and you get $13tn — more than the global PE market.
Apollo’s first funds would now be worth $74tn, just shy of global GDP. And Yale’s endowment? At a 36 per cent annual net return since 1990, its PE portfolio alone would be worth $5tn (the whole endowment is actually $41bn).
What’s going on? The widespread abuse of IRR, the private equity industry’s favourite but most hopelessly flawed measure of returns.
The Internal Rate of Return is a powerful tool — but it’s not a rate of return. It is a mathematical artefact. It assumes every dollar distributed is reinvested at the same rate it was earned. It’s not realistic, not additive, not comparable to market indices — and not designed to be used the way the industry uses it.
Apollo declined to comment, while KKR said:
We are proud of our long track record of investment excellence and confident we provide our investors with the requisite information to understand the performance of the strategies they are invested in. It is well understood how an IRR calculation is effected over time.
IRR’s most bizarre feature is that it becomes fixed early, and barely changes after that. That’s because of how it is calculated.
Here’s a simple example to show why. Let’s say you invest $30mn in 1976, get back $100mn in 1980 and then today what is left is worth $100mn. The annualised rate of return is only defined as what you have at the end over what you started with. But what you have at the end of 2024 obviously depends on what you did with the $100mn you received in 1980.
You earned a return on it that we do not know, and thus label x. So you have: (100+100(1+x)^44) You had invested $30mn, and so your annualised rate of return is: y=((100+100(1+x)^44)/30)^(1/48)-1.
But this is an equation with two unknowns, which is a bit of an issue . . . We don’t have unique solution to this. And so someone thought: “What if we set x=y? Then we have only one unknown and we’re good to go!” And this is IRR. A simple mathematical hack that become the dominant performance metric for a huge part of the global investment industry.
The complex mathematics mean that early cash flows dominate the calculation, while later ones have almost no impact. You can invest for 40 years, make or lose billions — and your since-inception IRR will still reflect that nice exit in 1980 or whenever.
In this example, the IRR is 35.1 per cent, a fairly common one in private equity. If today the investment would be worth $1bn instead of one $100mn the IRR would still be 35.1 per cent. Make it $10bn and it’s still 35.1 per cent. How about $100bn? Well now it goes up to 35.2 per cent! IRR is the most stubborn number in finance.
Why? Because if you treat IRR as a real return number the $100mn of 1980 is supposed to be worth . . . $56tn. So nothing that can happen today will change the overall return. You could bankrupt hundreds of billions of dollars' worth of investments in the meantime, but it wouldn’t matter to you, since your reinvested 1980 dividend is “worth” $56tn, circa half of global GDP.
This IRR stickiness is fertile ground for gaming. Fund managers know how to juice it. It’s not fraud — it’s strategy. But the consequences are real, as people are often presented these numbers as actual rates of return, and frequently believe them. This distorts the decisions of private equity firms.
For example, the fact that IRR is influenced by early distributions makes it possible for fund managers to strategically manipulate the metric by quickly exiting successful investments while retaining less profitable ones. They can also borrow to make their first investments rather than calling capital from investors.
Even without manipulation, IRR tends to be inflated — because of the survivorship factor. Firms that survived had good early exits (otherwise they wouldn’t have been able to raise follow-up funds) and their IRR is then set in stone. As a result, almost every firm that’s been around long enough claims a 20-30 per cent return.
This creates capital allocation distortions. It obviously favours private equity versus other asset classes, but it also distorts where money flows within the private equity industry, as the ability to manipulate it changes over geography and strategies. For example, funds in emerging markets or growth capital funds can’t boost IRR the same way the big US buyout funds can.
Most financial professionals believe they understand the issue at play here. They’ll recall that every textbook and finance course address the limitations of IRR as a performance metric. Unfortunately, this awareness leads many to assume they grasp the problem, and then to overlook it.
No textbook discusses the phenomenon of unrealistically large IRRs remaining large and constant over a 20-year period, despite large cash flows occurring in the meantime. That’s because textbooks focus on capital budgeting uses of IRR and do not consider the case of private equity. Consequently, most individuals assume they understand the issue, considering it basic knowledge, yet few can clearly and correctly explain the perverse patterns just described
Most private equity firms disclose different return metrics alongside their IRRs. For example, KKR’s 10K says that it has invested $203.4bn since is birth in 1976, and estimates that this had become $381.9bn by the end of 2024. Blackstone discloses its “multiple on invested capital” alongside its net IRR.
However, note that MOIC is gross of fees. So sadly, the most informative figure (net multiple) is not available. Note also how Blackstone did not have particularly high multiple in their early investments and their IRR is, as a result, one of the most modest ones out there.
Note that MOIC can also be manipulated by recycling capital, a trick not equally available across all types of private equity funds. Smaller fund managers are once again at a disadvantage here.
What can be done to eradicate the silliest of private equity performance metrics, or at least ameliorate its pernicious influence? Not much, but here are three ideas:
First, only report performance over rolling periods: 5, 10, 15, 20 years. These “horizon IRRs” — NAV-to-NAV IRRs — at least reflect real investment periods and reduce the distortion from ancient early exits. KKR’s 20-year IRR? Around 12 per cent. Yale’s? 11.5 per cent. Now we’re getting closer to reality.
Second, stop calling it a “rate of return.” Call it what it is: a discount rate that balances cash flows. “Internal Discount Rate”, or IDR, would be more honest — and more accurate.
Third, stop pretending numbers above 15 per cent mean something. We already report IRRs as “n.m.” (not meaningful) when they’re negative. For the exact same reason, high IRRs are not meaningful. They’re marketing.
But don’t hold your breath. IRR has become the theatre of private equity performance. It delivers a beautiful illusion. Until someone looks at the maths — or the cash. But given that this is by far the financial product with the highest margins, why would any advisor care to do so? Especially now that the multitrillion dollar US retirement market might forcibly opened up to private equity.
The industry insists that institutional investors are smart and sophisticated enough to see the limits of IRR and run their own numbers. Maybe. But it seems more doubtful that ordinary investors will see through the theatre. Tony Robbins certainly doesn’t.
Further reading:
— Is private equity actually worth it? (FTAV)
— The Tyranny of IRR (SSRN)
