Does “Cleverness” Lead to Superior Cash Returns? Venture Capital Cash Returns v Stocks and Bonds

There are a few people in the world who are lucky enough to be clever for their job – a job that pairs an excellent salary with extremely high upside.

Well-funded entrepreneurs, venture capitalists, private equity investors, and hedge fund managers are all put to work in the hopes that their cleverness will deliver outsized returns for limited partners (or investors in the case of entrepreneurs).

I often wonder just how much this cleverness actually translates into the success that is assumed, and if we are honest enough with ourselves about what financial success for truly clever people should look like.

For example, if you inherited $50m in 1975 and grew that to a $2B business empire in 2015…that sounds great. But if you had just dumped that into an S&P Index fund, you would actually have $3.5B, so in all your cleverness you destroyed $1.5B and made yourself illiquid.

If an entrepreneur raised $20m in 2009 and then sold his company for $40m in 2015, that earns him some serious praise for a big exit…except his investors would have been better off if he just took that money and dumped it in the S&P (and they weren’t remotely compensated for their illiquidity.) The entrepreneur gets credit for creating something new in the world (which matters), but he is a net destroyer of capital.

Sometimes I think I am just finding ways to keep myself busy by being an entrepreneur, when the most rational thing to do is probably just liquidate, put everything into Amazon, and let Jeff Bezos do the work for me. The issue is that humans are super bad at just sitting in rooms doing nothing, so we often invent all sorts of ways to keep ourselves busy being clever, and justifying it with potentially huge returns. Upside and the fun of building and creating keeps us going.

Along these lines, I was recently bantering with a VC friend about what I think is the over-glorification of the opinions of VC and Founder-gods. I think most people understand a couple narrow things incredibly well, and then otherwise they are kind of just saying things from their face (like I am doing right now, in this blog), and the world should be much less generally reverent.

The most potentially destructive opinion in the world is the one that comes from a smart, confident, charismatic person who is just bluntly extrapolating and pattern matching.

The crux of my irreverence argument is that unless someone has really shown that in all their cleverness they have actually done the basics of their job – beating the S&P and compensating for the illiquidity that comes with the clever deployment of capital (e.g. creating new companies, investing in private companies, taking long bets on a catalyst, locking up investors money, etc), then everything they say should be taken with a big grain of salt.

And further to that (and this is being a little picky), they should really be able to demonstrate they have done it in a statistically meaningful fashion that eliminates randomness and luck as the reason for their success.

So, I got into a bit of a research spiral about returns for venture capital firms to see how cleverness in this case actually played out in terms of cash returns, and thought I’d share my findings here.

 

Approach

Sources

Using these two sources, I compared the returns from Venture Capital to the returns from the S&P 500, 10-year US Treasury Bonds, and 3-Month US Treasury Bills.

1. US Venture Capital Index from Cambridge Associates
2. Annual Returns of Stocks, Bonds, and T-Bills from Professor Damodaran at NYU

Accounting for Illiquidity

I was also curious about how venture capital performed once you factored in an illiquidity premium (because it is worse for investors having their cash locked up with a VC than invested in liquid S&P index). I found 3% as an annual illiquidity premium from this Blackstone report, and that seemed reasonable to me, so I also ran some additional analysis, tacking on 3% to the S&P returns for every year to compensated for illiquidity, to give a rough baseline of what should be average VC performance.

Using Cash Returns Multiple as a True Metric for Performance

Because actual VC cash returns are entirely subject to a liquidity event (which can be wildly variable), looking at self-reported “paper returns” as a benchmark for performance felt too fluffy, so I wanted to examine the actual cash returns that VCs paid to limited partners. I used multiples instead of percentages because they are harder to manipulate.

(For the record, VC paper returns are sometimes stronger than the S&P…but until returns are liquid these private market valuations are just guesses/jazz hands.)

Allowing 10+ years for a Cash Return

VC investments legitimately take a long time to mature and funds are structured as ten-year investments (e.g. “it will take us a long time to get you your money, but when we do, it will be worth it”). So, for the purpose of comparison, I looked at VC funds from 1981-2006 to give VCs the necessary time to generate cash returns.

None of this analyis is going to be perfect and should be taking with a grain of salt (it is also very backward looking) – for example if a fund actually returned all its cash in 5 years, that isn’t accounted for. In an ideal world you would match cash returns for each fund with the market performance while that money was locked up.

Findings

• Even the better VCs (top 25th percentile) are not particularly good at making money for limited partners compared to more standard investments
  1. In 69% of years they lost to the S&P
  2. In 57% of years they lost to 10-year treasury bonds

• The worse VCs (bottom 75th percentile) are just terrible at making money
  1. Every year they lose to the S&P and 10-year treasury bonds
  2. In 88% of years they lose to 3-month T-bills

• “Good” venture capital firms almost never beat the benchmark returns of the S&P Index + 3% annual premium for illiquidity.
  1. In only 3 years out of 26 did the top 25th percentile firm beat this benchmark.
  2. In only 1 year out of 26 did the median firm beat this benchmark

• The top ~10% of VC firms must absolutely crush it. We don’t have data specific to this group, but mathematically, they have to be delivering outsized returns and propping up the entire asset class. This makes sense as otherwise there would be no reason for the industry to exist.

• Venture capital really benefited from dotcom 1.0, as the only years the industry truly outperformed the S&P + Illiquidity benchmark were 1994-1997. This does not bode well for the future of the asset class a whole, since dotcom 1.0 was pretty much fueled by VCs and investment banks colluding to loot retail investors. It is unclear how they can actually perform in a regulated environment that is “anti-looting”. Maybe the funds from dotcom 2.0 will be much better, but we just don’t have the data yet. My guess is that most returns for limited partners will be just as middling, since there is such a glut in capital and less opportunity for looting/passing the buck.

• VC is a bad investment unless you can get into the very top funds for a vintage, and as an overall asset class, it isn’t better than the S&P when you look at actual cash returned to limited partners.

So, at least by looking at past results, all venture capitalists are not created equal and their cleverness (as it relates to being a venture capitalist) should be revered accordingly. At the end of the day, just because people get paid to be brilliant and clever doesn’t mean they actually are (if you think these qualities should be backed up by clear results.)

As we look at the modern day crop of venture capitalists, it is probably worth keeping these statistics in mind.

There are a few great venture capitalists are legitimately amazing at their jobs, and we should all pay attention when they speak.

Most, and even “good” venture capitalists rarely create more value than clicking five buttons in your brokerage account and buying the S&P, so their general musings outside an area of demonstrated expertise are of unclear value.

And there are still plenty of bad venture capitalists who probably fall into the same category as the opinions of bad doctors. Well meaning, generally smart people who can probably kill you by accident.

————–

 

Here are some supporting tables from my data set if you want to check out more:

1) Even the good firms (top 25th percentile) rarely (3 years out of 26) match the S&P and compensate for illquidity:

Fund Vintage Year	Upper Quartile Return Multiple	S&P + Illiquidty Premium Multiple
1981	2.02	8.00
1982	2.20	6.77
1983	2.30	7.57
1984	2.04	8.72
1985	3.04	8.84
1986	2.07	9.55
1987	3.19	10.88
1988	2.52	8.55
1989	2.99	5.79
1990	2.80	4.70
1991	3.52	4.63
1992	3.18	4.77
1993	3.53	4.55
1994	4.43	5.18
1995	4.33	4.00
1996	4.60	2.12
1997	2.42	2.01
1998	1.56	1.80
1999	1.18	1.53
2000	1.22	1.93
2001	1.35	2.86
2002	1.21	4.12
2003	1.40	3.27
2004	1.14	2.88
2005	1.07	2.67
2006	1.12	2.25

2) The good firms are a lot better than the bad firms. On average the 25th percentile firm (2.4x) returned more than 2x the 75th percentile firm (1.12x).

Fund Vintage Year	Upper Quartile Return Multiple	Lower Quartile Multiple
1981	2.02	1.55
1982	2.20	1.54
1983	2.30	1.54
1984	2.04	1.27
1985	3.04	1.45
1986	2.07	1.46
1987	3.19	1.72
1988	2.52	1.45
1989	2.99	1.52
1990	2.80	1.60
1991	3.52	1.62
1992	3.18	1.55
1993	3.53	1.57
1994	4.43	1.35
1995	4.33	1.87
1996	4.60	1.27
1997	2.42	0.88
1998	1.56	0.62
1999	1.18	0.35
2000	1.22	0.44
2001	1.35	0.59
2002	1.21	0.48
2003	1.40	0.34
2004	1.14	0.36
2005	1.07	0.30
2006	1.12	0.31

3) The great firms are a lot better than everyone else. In 8 out of 26 years (particularly in the 1990s), the pooled return of all funds was higher than the top 25th percentile fund – that means the firms at the very high-end performed so well that they overcompensated for the bottom 75% of firms.

Fund Vintage Year	Pooled Return Multiple (All Funds)	Upper Quartile Return Multiple
1981	1.76	2.02
1982	1.79	2.20
1983	2.01	2.30
1984	1.77	2.04
1985	2.69	3.04
1986	2.90	2.07
1987	2.72	3.19
1988	2.42	2.52
1989	2.59	2.99
1990	3.15	2.80
1991	3.17	3.52
1992	3.09	3.18
1993	4.13	3.53
1994	5.40	4.43
1995	6.07	4.33
1996	4.89	4.60
1997	3.09	2.42
1998	1.48	1.56
1999	0.91	1.18
2000	0.92	1.22
2001	1.00	1.35
2002	0.80	1.21
2003	1.44	1.40
2004	1.12	1.14
2005	0.76	1.07
2006	0.85	1.12

4) The great firms also overcompensated for the bad firms (bottom 50%) – in every year but two, the average pooled return was higher than the median, which means the success at the top overpowered the lesser success at the bottom.

Fund Vintage Year	Pooled Return Multiple (All Funds)	Median Return Multiple
1981	1.76	1.90
1982	1.79	1.74
1983	2.01	1.85
1984	1.77	1.71
1985	2.69	2.38
1986	2.90	1.87
1987	2.72	2.28
1988	2.42	2.08
1989	2.59	2.09
1990	3.15	2.44
1991	3.17	2.36
1992	3.09	2.22
1993	4.13	2.30
1994	5.40	2.16
1995	6.07	2.68
1996	4.89	2.17
1997	3.09	1.35
1998	1.48	0.98
1999	0.91	0.76
2000	0.92	0.81
2001	1.00	0.83
2002	0.80	0.83
2003	1.44	0.67
2004	1.12	0.61
2005	0.76	0.64
2006	0.85	0.67

5) That said, it is hard to know who the great firms will be or get into their fund. Even if you knew that you were putting your money into the top 25th percentile firm, you were better off putting your money in an S&P index unless you invested in VC from 1991-1999. Once you compensate for illiquidity, you had to be invested from 1995-1997 to win your bet.

Fund Vintage Year	Upper Quartile Return Multiple	S&P + Illiquidty Premium Multiple	S&P 500 Return Multiple
1981	2.02	8.00	5.88
1982	2.20	6.77	4.95
1983	2.30	7.57	5.55
1984	2.04	8.72	6.42
1985	3.04	8.84	6.51
1986	2.07	9.55	7.05
1987	3.19	10.88	8.05
1988	2.52	8.55	6.29
1989	2.99	5.79	4.21
1990	2.80	4.70	3.39
1991	3.52	4.63	3.34
1992	3.18	4.77	3.44
1993	3.53	4.55	3.28
1994	4.43	5.18	3.75
1995	4.33	4.00	2.88
1996	4.60	2.12	1.49
1997	2.42	2.01	1.41
1998	1.56	1.80	1.26
1999	1.18	1.53	1.07
2000	1.22	1.93	1.36
2001	1.35	2.86	2.03
2002	1.21	4.12	2.96
2003	1.40	3.27	2.34
2004	1.14	2.88	2.11
2005	1.07	2.67	2.01
2006	1.12	2.25	1.74

6) Further, even if you knew you were putting your money into the top 25th percentile firm, you were better off putting your money in 10-year US Treasury Bonds unless you invested in VC from 1985-1997.

Fund Vintage Year	Upper Quartile Return Multiple	10-year Bonds Return Multiple
1981	2.02	4.50
1982	2.20	3.12
1983	2.30	3.73
1984	2.04	3.32
1985	3.04	2.91
1986	2.07	2.69
1987	3.19	2.60
1988	2.52	2.80
1989	2.99	2.51
1990	2.80	2.72
1991	3.52	2.37
1992	3.18	2.27
1993	3.53	2.04
1994	4.43	2.26
1995	4.33	2.02
1996	4.60	2.39
1997	2.42	1.93
1998	1.56	1.83
1999	1.18	2.31
2000	1.22	2.04
2001	1.35	1.75
2002	1.21	1.69
2003	1.40	1.70
2004	1.14	1.63
2005	1.07	1.58
2006	1.12	1.55

7) The bottom quartile of firms underperformed 3-month T-bills in 88% of years, which is pretty dark.

Fund Vintage Year	Lower Quartile Multiple	S&P 500 Return Multiple	3-month T-bills Return Multiple	10-year Bonds Return Multiple
1981	1.55	5.88	2.21	4.50
1982	1.54	4.95	2.07	3.12
1983	1.54	5.55	2.02	3.73
1984	1.27	6.42	1.93	3.32
1985	1.45	6.51	1.89	2.91
1986	1.46	7.05	1.87	2.69
1987	1.72	8.05	1.84	2.60
1988	1.45	6.29	1.83	2.80
1989	1.52	4.21	1.76	2.51
1990	1.60	3.39	1.66	2.72
1991	1.62	3.34	1.59	2.37
1992	1.55	3.44	1.56	2.27
1993	1.57	3.28	1.56	2.04
1994	1.35	3.75	1.57	2.26
1995	1.87	2.88	1.55	2.02
1996	1.27	1.49	1.50	2.39
1997	0.88	1.41	1.43	1.93
1998	0.62	1.26	1.37	1.83
1999	0.35	1.07	1.31	2.31
2000	0.44	1.36	1.24	2.04
2001	0.59	2.03	1.20	1.75
2002	0.48	2.96	1.18	1.69
2003	0.34	2.34	1.17	1.70
2004	0.36	2.11	1.15	1.63
2005	0.30	2.01	1.12	1.58
2006	0.31	1.74	1.07	1.55

8) The pooled returns of each vintage year (e.g. all the funds lumped together) lose to the S&P + Illiquidity Premium 84% of years (1994-1997), S&P 76% of years, 10-year bonds 57% of years, and T-bills 42% of years.

Fund Vintage Year	Pooled Return Multiple (All Funds)	S&P + Illiquidty Premium	S&P 500 Return Multiple	3-month T-bills Return Multiple	10-year Bonds Return Multiple
1981	1.76	8.00	5.88	2.21	4.50
1982	1.79	6.77	4.95	2.07	3.12
1983	2.01	7.57	5.55	2.02	3.73
1984	1.77	8.72	6.42	1.93	3.32
1985	2.69	8.84	6.51	1.89	2.91
1986	2.90	9.55	7.05	1.87	2.69
1987	2.72	10.88	8.05	1.84	2.60
1988	2.42	8.55	6.29	1.83	2.80
1989	2.59	5.79	4.21	1.76	2.51
1990	3.15	4.70	3.39	1.66	2.72
1991	3.17	4.63	3.34	1.59	2.37
1992	3.09	4.77	3.44	1.56	2.27
1993	4.13	4.55	3.28	1.56	2.04
1994	5.40	5.18	3.75	1.57	2.26
1995	6.07	4.00	2.88	1.55	2.02
1996	4.89	2.12	1.49	1.50	2.39
1997	3.09	2.01	1.41	1.43	1.93
1998	1.48	1.80	1.26	1.37	1.83
1999	0.91	1.53	1.07	1.31	2.31
2000	0.92	1.93	1.36	1.24	2.04
2001	1.00	2.86	2.03	1.20	1.75
2002	0.80	4.12	2.96	1.18	1.69
2003	1.44	3.27	2.34	1.17	1.70
2004	1.12	2.88	2.11	1.15	1.63
2005	0.76	2.67	2.01	1.12	1.58
2006	0.85	2.25	1.74	1.07	1.55

Here is the full data in excel format set if you want to play around with it – if you see something wrong with my inputs, please let me know: https://www.dropbox.com/s/3kj2p2005nmyiej/Venture%20Capital%20Cash%20Returns%20Analysis.xls?dl=0

_______________________________________

Notes about tables

1. A multiple of 2.0 in these table means that for every $1.00 invested, $2.00 was returned in cash.
2. The stock and bond multiples reflect a 12-year period from the year of initial investment, in order to best mimic the total amount of time it may take a VC firm to return capital (most VC funds take more than 10 years to wind up). For years 2004-2006, the multiple reflects prices at the end of 2015.
3. The VC multiples reflect cash returned up to March 31, 2016, but the stock and bond multiples reflect cash returned up to December 31, 2015.
4. The VC multiples reflect a multiple on “paid-in capital” which could happen any time over a fund, whereas stock and bond multiples assume all cash is paid upfront. I actually think the comparison is still fair, since capital is committed upfront either way.
5. I don’t have information about how the very best VC firms performed, the most visibility I can get is top quartile but the numbers indicate that they really pull weight. My guess is that the top 5% of VCs earned ~75% of returns for the industry as a whole.