Big Boys Returns

Show me your net multiple

People keep asking me for an update on the Billionaire Factory. Fine. But let’s do it a bit differently this time (7 years on).

In that paper, which remains the most read thing I have ever written, I did something deliberately old fashioned: I took the gross multiples straight out of SEC filings, so that anyone could check my work. I then imputed a theoretical carry. I did not know the other fees, nobody does, so those were left out. Even with that generous treatment, the multiples were not great.

And I noted something in passing, almost as an aside. Everyone was busy posting net IRRs and gross IRRs. The multiple, however, came only in gross form. Since the only multiple that means anything to an investor is the net one, that was, let us say, a bit of a pity.

So this time I went to Preqin and took the most recent data.

The process

Nothing exotic:

• Preqin, all funds

• Vintage 2020 or earlier

• Only funds with a reported net multiple

• Final close size reported at $100m or above

• Net multiple taken in the currency in which it is reported

• Weighted by fund size in USD as given

That is it. No cherry picking, no adjustment, no clever normalization. I take what the database gives me.

One thing worth saying out loud. To use a net multiple you have to trust the NAV. NAVs usually do not reflect the carry still to be charged. A NAV is simply the sum of what all the current equity is worth before carry, at supposed fair value. So it is not a hard number. But it is at least the number investors actually observe, which is more than can be said for a gross multiple.

The results

Same top four as in the paper. The four that hoover up essentially all the capital.

Firm Capital in the calculation Size weighted net multiple

Apollo $92bn 1.39x

Blackstone $223bn 1.59x

Carlyle $70bn 1.55x

KKR $80bn 1.79x

We are told constantly that private equity has massive dispersion across funds, and that this is precisely why manager selection is the whole game, why you must pay for access, why the big names are worth the queue. Where exactly do you see that dispersion? These are the largest managers on earth, running very different strategies on paper, and they all land within a whisker of each other.

While we are here. Everybody enjoys making fun of CalPERS for being terrible at private equity. Their net multiple? 1.5x. Of course. They invest in all of these big guys, so they get the returns of these big guys. CalPERS reports an IRR of 11%, and an IRR somewhere between roughly 0% and 15% is actually meaningful, so that checks out too. Simple, one would think.

Oh and would be great one day to know the net multiple of Yale, you know, the model! Am willing to bet it is close to 1.5x too. Pity they never reported it, only IRRs. Was enough to make them a model apparently…

What 1.5x actually means

Let us assume an effective holding period of five years.

Firm Net multiple Implied annual rate over 5 years

Apollo 1.39x 6.8%

Blackstone 1.59x 9.7%

Carlyle 1.55x 9.2%

KKR 1.79x 12.3%

And remember what period this is. Interest rates were zero. Equity prices went up and up. This is what a leveraged equity strategy delivered in the friendliest environment it will ever see. US stock returns over the last hundred years are like 10%-12%, it is a bit less for the whole world, like a bit below 10%.

Should it have been higher return for this levered equity strategy then? Yes, obviously. But as I pointed out years ago, if you charge something like 7% a year in fees, a great deal is leaving the table on the way out. The money did not vanish. The private equity fund managers at these big firms bought very large houses and planes. They did very well, thank you. It all checks out.

So next time somebody is visibly excited about investing in a big name because the big name makes so much money, just ask: what makes you think that? Chances are the answer is the flashy and meaningless IRR. And when they tell you that of course nobody looks at an IRR without the net multiple next to it, then fine. Show me those net multiples.

Meanwhile you have gazillion of hard working smaller managers, especially in frontier market and they have, I kid you not, on average a multiple of 1.5x, but everybody is convinced that the big names are the people who returned the most, by far, of course. Yeah, on an IRR basis, they did.

I hear you

You may be thinking: hold on, but KKR is at 1.79x, clearly they are different, clearly the big name premium is real.

Yes. Of course. About that.

Preqin does not manage to find a net multiple anywhere for a large chunk of KKR’s funds. Isn’t that beautiful? The most meaningful performance metric, the one that investors are supposedly focused on, is missing…

KKR’s total fund size in Preqin is $173bn. The calculation above includes $80bn. So roughly half the money is missing, because the net multiple is missing.

Is the missing half a random sample? Who knows?

But, among the KKR funds that do have a net multiple, the mean IRR is 11%. Among those that do not, it is 9%. But you cant average IRR, so…

Ah yes, because the IRR is always reported for KKR funds. The IRR is never missing. The IRR is never shy.

Caveats

I am using Preqin data. I take their net multiple in the currency it is given, weight by fund size in USD as given, and include every fund in the dataset that meets the screen. That is the whole method. If Preqin is wrong, I am wrong.

But here is the thing. I do not have to be right. The firms have the numbers. All four of them know their own size weighted net multiple to the third decimal, across every fund, in every currency, since inception.

They could publish it tomorrow.

They publish the IRR instead. And a multiple but gross of fees. Of course.