In simpler times, the usual story was: there's a fund manager. They manage a fund. The investments they make are through that fund, and they collect fees from these investors on a standard schedule. But things have often been messier. Consider Warren Buffett's early structures: he formed separate partnerships for different investors, and let them opt in to a mix of fees and hurdle rates. Some of them paid him a sixth of all profits, some paid a third of everything over 6%, and his first group of investors paid him half of all profits beyond 4%. Which was not the end of the complexity; at first, Buffett also had a substantial personal account, which he used for trades that didn't scale as well as those in his main vehicle. At the beginning of 1962, he consolidated everything into one fund with a 4% hurdle and 25% fee, and put most of his liquid net worth into it.¹ Investment vehicles aren't as ad hoc as they were when the concept of a hedge fund was being independently reinvented by people reviewing the details of the Investment Company Act of 1940.

Markets have evolved since then, and a big part of that evolution is recognizing that strategies' capacity is on a continuum: some strategies require a minimal amount capital, but can't scale forever (these are the kinds of things Buffett presumably did in his PA, and thus the source for charming but non-scalable stories like buying one share of a duck-hunting club turned oil royalty trust). Some strategies scale a lot, and need that scale if they're going to produce meaningful returns. And then in the middle, there are strategies that have variable capital needs, or that might evolve towards being able to absorb more capital.

So, some investment management firms will have multiple vehicles, with different fee structures, to capture value from the same means of alpha production. This is basically the story of multi-manager hedge funds: their founders often started out doing something that could produce high returns on a limited capital base but couldn't 10x those returns by 10xing the capital. But those strategies were adjacent to strategies that could absorb more capital. Renaissance Technologies is the classic case: they started out managing outside capital, and got good enough returns that they could buy out their external investors in the early 2000s.² And then, a few years later, they launched a separate fund, Renaissance Institutional Equities, with more capacity than their main fund. It hasn't performed remotely like Medallion, but they never pitched it that way—the basic pitch was that their miscellaneous leftover signals were still pretty good, even if they didn't hit the firm's internal return targets.

That's a pitch that gets easier and fairer to make over time, because it neatly solves an adverse-selection problem: if you're pitching a fund that does 50% annualized returns and never has a down year, the first question isn't about process or repeatability, it's "Why do you need my money? And how could it be worth it to tell us, or anyone, how much money you're making?" Having separate vehicles solves that problem: the good stuff is so good that you can't afford it, but the less-good stuff is still pretty good.³

But this still means that the fund manager is taking some risks, and not with their own money. Suppose a fund looks at the various alphas it's been able to generate and notices a distinct pattern: they're harder to find and they don't last as long as they used to. And they also notice that it's hard to accelerate their research fast enough to offset this. One thing they could conclude is that they used to be able to make a lot of money, and soon the money will have been made. But the other thing they can do is ask themselves whether they could package whatever working signals they have into a fund that will have a great first-year track record, look decent in year three, and still be able to raise money in year five, after which their focus turns to management fees rather than performance fees. This would be bad news (and it puts allocators in the strange position of hoping that whoever they're dealing with is making absolutely monstrous returns in their core, unavailable-to-outsiders strategy, year after year, so there's enough excess return to go around).

This gets more acute when you consider the case of private equity continuation funds, especially the new variety that charge higher fees. The basic idea of a continuation fund is that the period in which a PE fund operates—typically, ten years plus a two-year extension—doesn't necessarily match the period during which it's a good idea to own a particular asset. So, towards the end of the fund's life, the manager sometimes proposes a new vehicle that will raise money to buy some specific asset or collection of assets and keep running it. The two most common proposed motivations for this are:

1. It's hard for PE firms to find good companies to invest in, so they'll do anything for fees, including holding on to a company for years longer, or

2. It's hard for PE firms to find good companies to invest in, so when they find one, they should basically never let anyone else own it unless they get a ludicrously generous offer.

It's probably a mix of both.⁴ But the picture gets complicated when a company owned by a typical-fee vehicle is being sold to a higher-fee vehicle. There are a few drivers here: the actual change in fees redistributes some future upside to fund managers from their limited partners, and the lower the price at which the appreciated company is sold to a new fund, the higher those fees are. On the other hand, those fees get realized ten years in the future, whereas the carry from the transaction happens right when the new vehicle buys it. What's more powerful is just owning an appreciating asset, and collecting fees on it, over a longer period. So LPs have to play a little game where they try to figure out exactly what the GP is getting out of this transaction: do they want upfront liquidity from realizing a gain, or do they actually want to maximize their equity stake, through direct investments plus fees that are equivalent to a call option on 20% of the remaining equity.

The upshot of all of this is that there's no free lunch: picking the right hedge fund or PE firm should be about as hard as picking the right stock or acquisition target, with the only difference being that your counterparty's sophistication leans towards the financial side rather than operations.

The question of who charges what kinds of fees has been a running theme in The Diff for a while. We've looked at:

• What makes Rentech special.

• What to do with a strategy that works but doesn't scale (find another one!).

• Hedge funds have slowly renegotiated their deal with outside investors ($).

• And PE firms have started holding some assets indefinitely ($).

Join the discussion!

1  This is the kind of thing you want to do before you have to, rather than sheepishly after a year where your fund underperforms the market or loses value and LPs who know about your personal trading ask "And how did you do this year?" So, it may not be a coincidence that he did this right after 1961, when the partnerships' returns were +46%, their best annual result up to that point.

2  They still had passive capital from retired employees, which constituted a kind of retirement plan mixed with a noncompete: there are very few jobs that pay quite as well as keeping eight figures in Medallion Fund for a few years until the stake gets redeemed to make room for an investment from someone who's still working there.

3  It's close to the logic behind celebrity chefs offering branded foods and ingredients. The people who can't dine at their restaurants quite as often as they'd like still get something chef-endorsed and reasonably good, at a price point that's higher than comparable products but lower than the cost of the real thing.

4  Or a third more cynical reason: no one wants to buy the asset(s) in question just yet, but if you wait long enough a bull market in that kind of asset might bail you out.