# How Hedge Funds Get Rich (Hint: It’s Not Their Returns)

I used to think that those that ran hedge funds got rich because of their incredible returns.  Then I heard about the 2 and 20 fee structure that most hedge funds charged.  The typical hedge fund fee structure (historically) is 2% of assets under management and 20% of all positive returns.  Therefore, if you gave a hedge fund \$1 million and they got a 10% return on it, their total take in fees would be:

[\$1 million * 0.02] + [(\$1 million * 0.1) * 0.2] = \$20,000 + \$20,000 = \$40,000.

This represents \$20,000 for managing your money and \$20,000 for their positive performance (i.e. the \$100,000 they earned with your money).  Despite their \$40,000 fee, you would end up \$60,000 richer than before you met them.  Good deal, right?

However, what if I told you that in less than 20 years the hedge fund would have more money than you (regardless of the size of your initial investment).  Shocking, but true.

I created a simulation model in R that illustrates this (program is here).  It assumes that the hedge fund will receive a market return each year (~10% average with a 20% standard deviation) and reinvest all fees it gets along with its clients’ money.  After running 10,000 simulations of this model, the average amount of time before the hedge fund’s capital exceeds the client capital is about 17 years.  You can see the hedge fund’s capital as a percentage of the client capital for the first 100 simulations here:

The horizontal black line represents the point at which the hedge fund has as much money as its clients.  The dashed gray vertical line is the point where the hedge fund, on average, would have more capital than its clients.  You can see that most of the 100 simulations cross the 100% line before the 20 year mark.  The luckiest hedge fund is richer than its clients within 13 years and the unluckiest within 23 years.

Now, compare this to a lower cost ETF/index fund option (see plot below).  If the client were to purchase an S&P 500 ETF/index fund from Vanguard or Charles Schwab with an annual expense ratio of 5 basis points (0.05%), the time it would take for the fund to have more capital than its client is almost 1,500 years!  Remember, this is the exact same simulation as above, but with lower fees.

The stark contrast between the hedge fund fee structure and a low cost ETF should illustrate how extractive the 2 and 20 hedge fund fee structure is for investors.  This is true even when the hedge fund has a watermark (a performance benchmark it must hit before it gets a performance fee) or whether the hedge fund can outperform the market.  I ran sensitivities on both of these factors and they made little difference.  No matter what initial capital you give the hedge fund to start with, the hedge fund will become richer than you since its real talent is transferring your wealth into its coffers.

The good news is it seems that investors have caught on to this fact as hedge funds saw increased outflows in 2016 even as they dropped their fees.  However, the current fee averages (1.4% management and 17% performance) don’t change the math enough to make much of a difference.  With these fees, a hedge fund would have more capital than its client within 22 years instead of 17.  These are both a far cry from the ~1,500 years that would be required by a low cost index fund/ETF.

How You Can Get Rich (Hint:  Pay Lower Fees)

The moral of the story is that fees matter, a lot.  When it comes to investing, a little more in fees can go a long way (for those charging them).  So though 2% (or 1%) might sound small now, within a short time frame it can add up very quickly.  I recommend keeping your average fees below 0.1% across your portfolio.

This is post 01.  Any and all code I have related to this post can be found here with the same numbering:  https://github.com/nmaggiulli/of-dollars-and-data

Disclaimer

The above references an opinion and is for information purposes only.  It is not intended to be investment advice.  Seek a duly licensed professional for investment advice.

## 16 thoughts on “How Hedge Funds Get Rich (Hint: It’s Not Their Returns)”

1. alephblog

I got an average of 23 years using a return distribution N(.1,.2) as you did, and 2% and 20%, but imposing a high water mark provision, which is pretty standard, you might not have done. I can post my code to the web — but it is an Excel spreadsheet…

Nice start to your blogging… keep it up.

Like

• My program does have a watermark option and I did run a few of these simulations (not posted here), but just like you found, the main conclusion did not change much. If I remember correctly, using a 5% watermark had an average of something like 22 years, so your Excel model is a great sanity check. Thank you!

Like

2. Fatman

Performance fee is usually calculated on the net return. (this is overly simplified, as management fee payments and HWM payments have different periodicity)

\$Return = %Return * AUM
\$Management_fee = X% of Assets under management (AUM)
\$Performance_fee = Y% of ( \$Return – \$Management_fee)

There is always a High Water Mark (HWM), calculated on the same dates as the Performance_fee.
There is usually a carry forward on HWM, that is if the fund has lost money ( net of all costs) and the client redeems some of the AUM
they can reinvest ( usually within 2-3 years) and recapture that HWM ( ie no free lunch for the hedge fund).

And big investors often have a claw forward clause, ie if you make say 10% on 100mm in the first 6 months they have the right to up the numbers
to say 200mm for the second 6 months, if the fund loses 5% in the second 6 months the performance fee is zero even though the return on AUM is 5%.

2/20 is not the norm. it is the norm for some famous old school hedge funds ( household names).
These guys have amazing historic performance but are much the same as the majority of the industry in recent years thus the market noise about fee.

There is a wide range of fee structures but your average HF with a 12% ‘ish vol fund is charging something like 1.2/20, or 0/30.
There should be no passthrough fees other than those required to run the fund ( ie Fund Admin, Audit, Trustee etc).
The costs of the hedge fund manager should 100% be eaten by the hedge fund manager.

The majority of hedge fund managers have less than 100mm under management and are just paying the bills, often not paying the partners at all.
You need about 500mm+ to make a reasonable return on your effort. Some guys can do it with 200mm but they are being really lean.
Magically 500mm is the lower AUM hurdle of most US Pension plans, they will not invest ( even through a managed account) with a manager that has less than 500mm. Interestingly, the strategies that make a good return are often capacity constrained thus the 500mm hurdfle dooms Pensions to investing in low (risk adjusted ) return managers thus deepening the short fall.

I look forward to seeing you post the results for the next version of your model.

Like

• Thank you for the in-depth analysis on this simulation. I appreciate the rigor and time you put into this. I agree that this is a very simple model in many ways. I can try some of the sensitivities you noted and provide the results in a comment. Give me a few days and I will get back to you.

Like

• Dear “Fatman”,

I did many, but not all, of the sensitivities you requested and they make no material difference for my conclusion. I will provide the results for each sensitivity as I layer them on. The number I report below is the point at which the hedge fund has more money than its client in 50% of simulations:

1. Hedge fund 2% management fee and 20% performance fee (base): 17 years (this is what I show in the post above)
2. Hedge fund 2% management fee and 20% performance fee net of management fee: 18 years
3. Hedge fund 2% management fee and 20% performance fee net of management fee + 8% watermark: 19 years
4. Hedge fund 1.2% management fee and 20% performance fee net of management fee + 8% watermark: 23 years (you suggested this sensitivity)
5. Hedge fund 0% management fee and 30% performance fee net of management fee + 8% watermark: 22 years (you suggested this sensitivity)
6. Hedge fund 0% management fee and 25% performance fee net of management fee + 6% watermark: 26 years (this was Warren Buffet’s fees in his early days)

As you can see none of these even come close to the ~1,500 years it takes for a low cost ETF/index fund to have more wealth than its clients. I apologize for not modelling the clawback provision, but I didn’t want to spend the time creating it when my intuition is telling me that it won’t materially affect my conclusion either.

I have no agenda here. This is all mathematics playing itself out. You can’t charge 20-40 times more in fees and expect it to turn out well.

Like

3. John

This is very interesting. Thank you for starting this Blog.

Like

4. Darren Sinden

How many hedge funds have a longevity of 13 to 23 years or more ? secondly if they achieved that tenure what are their returns over that period compared to the index tracker fund ? Because it seems to me that the natural extension of your argument and data (which is v interesting btw) is that you should settle for Beta IE the market return ( mediocrity?) rather than pay a high price for potential out performance that the hedge fund may offer. (I think thats what Bernstein had in mind with their passive investing is worse than Communism comments) Great news for unsophisticated investors and institutions without an edge in picking managers or the expertise / scale to do so and this echoed to some extent in the recent Credit Suisse paper looking for easy games
But many HF investors don’t fall into that category .

My other comment is this if you take your line of thinking further then if you were CEO of a sales led company would you start to resent the cheques you wrote to your best salespeople rather than being pleased because they were earning more than you ?

Like

• Thank you for your reply. You are right that many hedge funds don’t have longevity of 13-23+ years, but I don’t think that changes my point. They don’t have that longevity for a variety of reasons not covered here. You are also right that those that can last that long will have likely outperformed the index after fees, and I am not against those funds. They deserve to get rich if they are truly providing extra return (alpha). The problem is every hedge fund thinks they are that fund. Historically, most hedge funds have not outperformed the market after fees over extended periods of time.

As for the mediocrity argument I am saying “yes!”. It is far better to get the average return each year and pay almost no fees (0.05%) than to try and beat the market and pay for it big time (through high fees) if you fail to achieve this goal. This is more a question of risk vs. reward. And personally, I don’t think their fees currently justify the extra possibility of reward. This is just my opinion. An investor with a larger appetite for risk would disagree with me.

If I was the CEO of a sales led company I would not resent the checks I wrote to my best sales people…unless they stopped selling so well.

Like

5. Amit Chokshi

I would be curious if you ran a sim based on no mgmt fee and hurdle rates vs 2/20 and HWMs. The mgmt fee + 20% is the killer. Even with a 10% return, why pay 4% in fees for what ultimately is likely to be a risk profile comparable to a low cost 60/40 and is likely less tax efficient? Secondly, risk and return do not correlate 1:1 so the net return delivered by a manager must account for the 3-5% of additional fees which may or may not correspond to an additional 3-5% of risk. In today’s low return market, just look at the vol for instruments trying to deliver 6-8% vs 2%, the 400 – 600 of additional bps through return seems to correspond with double or triple the corresponding risk.

The ideal HF comp model is no mgmt fee and a 20% take over 6% hurdle. That seems to at least align net fees with performance. In your scenario a 10% gross return equates to 9% for the investor and 1% for the mgr. An 8% gross is a 7.5% net, which while still about 10-15bps for index costs, is only 30-40bps higher.

Like

• Thank you for your response. If you look in one of the comments above I posted a few sensitivities based on another reader’s comments. In that comment I did a 0% management fee, 25% performance fee with a 6% hurdle (I called it watermark, but it is really a hurdle) and the number of years when 50% of simulations had more money than the client was 26. This is definitely higher than the 17 years for my base 2 and 20 model with no hurdle. I remember doing one with 0% management, 20% performance with a 6% hurdle and the number of years there was ~33 if my memory serves me correctly, but I did not post this. Yes it is longer, but it is still nowhere close to the 1,477 years for the low cost index fund/ETF. Hope this answers your question.

Like

• Amit Chokshi

Wow, that intuitively does not make sense to me re hurdle rate leading to more transfer of wealth to GP from LP v 2/20. If I am doing basic math here, 2/20 assume \$100 and gross return of \$120. The 2% fee against avg capital over the year of \$110 would be \$2.20. So deduct from \$120 u have \$117.8, take the 20% of the \$7.78 profit and the LP is left with \$114.24. In the same scenario with a 6% hurdle, 0% mgmt fee, and 25% performance fee, there is \$0 mgmt fee and the GP gets to take 25% of a \$14 profit (100 x 1.06 = 106, then 120 – 106) or \$3.50 leaving the investor with \$116.50.

The lower the returns the worse off an investor is with 2/20, if u get a 5% gross return, the net return to the LP is like \$102.5 in 2/20. With a 6% hurdle, the investor keeps 5%.

The only way it seems the 6% hurlde does better is at the very high returns, it seems even 50% returns in a year leads to a better end result for an LP under a hurdle rate…

What am I missing?

EDIT (from OfDollarsAndData): I can’t add more comments so here is my response:

No your intuition is correct. The hurdle makes the transfer of wealth slower. It takes more time for the GP’s money to be greater than the LP’s money when you introduce a hurdle. I am sorry if that was not clear. With no management fee and a 20% performance (w/ 6% hurdle) it takes 33 years which is much longer than 17 years. This is good for the LP as their fees are lower.

Like

6. Kevin

Thanks for your post. I always felt that fees weren’t discussed enough, and as an investment analyst, I always tell everyone I know to crunch the numbers on their fees before anything else. You don’t mention investment skill much – some managers really do earn their fees, but the 2 and 20 fee structure is a relic and I don’t think anyone is going to do really well for clients with that structure long run (back in the day, being L/S you actually earned a healthy interest income on your short book, while today you’re paying to borrow).

Anyway, maybe I’m a little daft, but I don’t see how mathematically its possible for a GP to overcome an LP in dollars (from a standing start), when charging no mgmt fee and only a performance fee with a hurdle rate? By definition, if you charge only a performance fee and invest alongside that same investor, they will always be ahead (unless the performance fee is north of 50%)? Am I missing something?

Like

• It is possible for this to happen with no management fee because the GP’s money grows faster than the LP’s money. Remember, I assume the GP reinvests its money along with its clients. The GP’s money keeps growing at 10% (average) annually while the LP’s money grows at (10% * (1 – performance fee)) annually on average. As long as the performance fee is above 0%, the GP will eventually catch and pass the LP with enough time. That is the simplified form. Technically the GP is also getting the performance fee adding on to their capital base every year. So they are growing at more than 10% per year.

If I drive my car at 100mph and you are driving at 80mph, no matter how far you are ahead of me, I will pass you eventually. The question of when I pass you is the point of this particular post.

Like

• OK fair enough, that makes sense. By the way, I did a quick excel using 0% mgmt fee, 20% performance fee, 6% hurdle and 10% per annum returns…it took 78 years for the GP to catch the LP in mine (my math may be wrong, but pretty sure I captured the assumptions).

I do appreciate the discussion of fees and the concept you’re illustrating. I should say though that under the fee structure I just mentioned, talking about how many years it takes the GP to catch the LP really misses the point. Investment mgmt is like anything else – there’s what you pay and there’s what you get. A performance fee above a 6% hurdle with a high water mark and in which the GP is also fully invested is a fee structure that fully aligns GP and LP. If the GP does well under that structure over many years, than it’s because they’ve earned their fee.

I like to think of fees in a multi-step way. First, is your manager beating his benchmark over long periods of time NET? (make sure benchmark is appropriate).

Second, what % of gross return are you keeping vs what you’re giving up? In the structure I referenced, the investors will keep ~85-100% of the gross returns in most years. Whereas there’s many equity funds charging 1% on assets, and earning 5%, i.e. investors are keeping 80% for sub-market performance.

Anyways, thanks for the post, fees are important!

Like

• Kevin,

There is no way for me to check your math and I don’t think it matters all that much. 78 years is not near 1,500 which is the point.

And yes you are right. If the GP outperformed the market by more than their fees every year then you would be correct. That is a big if though. Using the fee structure you suggest, if the market gets 10% a year, a hedge fund would have to outperform the market by a little less than 2% a year to be the same as a low cost index fund. Beating the market is one thing, but consistent outperformance of the S&P 500 of over 2% year after year is quite rare.

As for your remark about equity funds charging 1% a year, I am against those just as much as the hedge funds with high fees. I have nothing particular against hedge funds, I think their fees for most of them are too high to justify their value.