Andreessen Horowitz has a very strong spray and pray feel to it relative to other funds in its class (much like YC). Look at how enormous its portfolio page is: https://a16z.com/portfolio/. I bet this list isn't all inclusive, either.
For comparison: Founders Fund has an IRR of ~55%, at ~$1B AUM scale. It tends to invest in fewer companies -- with much higher bar and conviction -- and its portfolio has a much lower failure rate than competing funds (of course failure rate doesn't matter as much for VC returns, but it's still an interesting fact).
I have no connection to FF whatsoever, but have learned a lot from the way they invest and much prefer their model to the spray and pray style (YC, Ron Conway, A16Z, etc).
YC (and probably Ron Conway) are apples to oranges with a16z. Spray and pray works when you are looking at 10000x multiples on your best investments. It doesn't when you are deploying billions and don't come near that order of magnitude for your best bets.
As far as I know YC doesn't have a single 10,000x investment (i.e., $100B+ exit). Not one. Yet I once listened to a YC video where Michael Seibel (President of YC) discussed some of their stats. He said they've funded over 2,000 companies, and of those have 17 unicorns that are worth ~$100B in aggregate valuation. So that means their hit rate is generously
17/2,000 = 0.85%
Compare to i.e. Jason Calacanis who on his own has a hit rate of better than 1 in 20. Now assuming YC paid $100K per company and gets to keep a blended 1% of the $100B (is that too small?), they've put in about $200M in funding to get back
$100B*1% = $1B
to net roughly $800M in profit for their stakeholders. So they're a 5x fund. But that's really...not that good...(at least it's not world class).
But am I missing something? They've definitely gotten a lot better at picking companies during the Sam Altman era (by, IMO, funding deep tech companies that actually have the chance of 10,000xing), but it'll still take another 5-10 years to really prove that.
Now YC might argue that they're not purely a profit-driven fund. And that's true. But isn't it a bit worrying that after thousands of investments they haven't funded a single $100B+ company? YC has an enormous influence on the startup ecosystem. Is an institution with a 0.85% hit rate really sending us the right lessons?
Interesting points. I'd say creative destruction is as volatile as it gets and 1% might be pretty large. On 95% of this planet the odds of creating a unicorn are about zero. Questions: How many 100B+ firms are there? How many arrived in the 'old economy'? How many failures for each ramen, for each SME, ..., for each unicorn?
In my opinion it's enormously unusual to find a product market where 100B in future profits are up for grabs and a testament to mans inventiveness that we are even having this discussion. In a competitive market I'd expect the chance of a unicorn to approach zero. And it doesn't! It's awesome and takes a whole lot of failure.
On the side of the investors I'd be in the boat thinking it's more of a lottery than a skill, but these funds seem to prove otherwise so while I wouldn't invest in them and caution my company to be careful, I hope my pension fund is in them, a little.
Few comments:
* As rightly pointed out below, YC's investment used to be $20k for 7%, meaning they had a 10000x return on AirBnb and Stripe (so far)
* YC has funded increasing numbers of companies over time. Of the 2000+ companies that have been funded, the majority (75%) have not reached maturity.
* Yes there is dilution to the 7% stake, but it is not as significant as mentioned, especially for the most successful companies. I would estimate their stake at closer to ~3% after dilution, not 1%.
Take all these together and you are looking at a much higher fund multiple than 5x. My estimate would be at least an order of magnitude higher.
Is there a fund that is better than 5x in aggregate? Everybody flaunts their unicorns, but at the end, those bets are balanced by many more losing bets...
Don't forget back when YC invested in AirBnB and Stripe that they only put in $20,000 for the 7%, which brings down your valuation quite a bit closer to something that might include AirBnB and Stripe as 10,000x exits.
As I understand it, YC has the right to participate in priced rounds at the same valuation as other investors to retain their 7% stake. That does reduce the return on their most successful investments (since the denominator is bigger), but it also causes those investments to make up a larger share of their portfolio, which helps their aggregate ROI.
For people who want to dig into Founders Fund's philosophy:
FF's Bold bet:
- Space X: Invest 10% of its fund in 2008. - Stemcentrx: Invest $300 million on Stemcentrx. AbbVie acquired Stemcentrx for up to $10.2 billion. Founders Fund owned about 16%.
> ...with much higher bar and conviction.
Agree. Here is a quote for it.
“The key to the strategy once we have conviction we are willing to invest a lot. So with Stemcentrx, over the multiple rounds of this company, we invested something like $300 million. It’s not enough to think something is going to be one of the most important companies on the planet you have to back it up. So, to me, venture capital is about having conviction and putting a lot of money behind it.”
- Brian Singerman(GP of FF)
> Look at how enormous its portfolio page is...
> It tends to invest in fewer companies -- with much higher bar and conviction
I created two theories[1] to explain the difference in investing perspectives between Marc Andreessen and Peter Thiel:
(a16z ≠ Andreessen, Founders Fund ≠ Thiel)
Run Faster vs. Jump Higher
Higher: Investing for Control
You can term Peter’s approach to entrepreneurial strategy: investing for control.
You achieve it by patiently building your business at the same time as ensuring that you will be insulated from future competition. It takes time and it takes investment dollars as resources. Put out some crappy minimum viable product and you lose some options to control because you have shown your hand to others. Instead, what you want to do is put out a complete product with a strategy to acquire the complementary resources to insulate it from future competition.
There is, however, another sort of monopoly — a path that gets you 100 percent of a real market. This is done by having the capabilities to beat all rivals on either quality or cost. To see how this arises consider a very structurally price-competitive market (in economics, Bertrand competition). Now suppose that you develop an innovation that allows having lower marginal costs than everyone else. In this situation, you will be able to capture 100 percent of the market and so you will be a technical monopoly.
Faster: Focusing on finding the timing
You can term Marc’s approach to entrepreneurial strategy: focusing on execution(finding the timing).
It is not so obvious that one path to monopoly is more profitable than the other. If you focus on execution you can get to market quicker and with fewer resources. You can learn as you go and actually invest for the capabilities that will give you a competitive advantage in the future. In other words, while it takes on-going work — no resting on your monopoly laurels here — you can still ‘own’ a market. The difference is that your pricing is constrained by potential competition from other firms.
Finding the timing is the most important part of the execution. You need to execute your idea fast and good enough within the critical window to succeed. Most importantly, you may need to survive long enough to find the right window.
So, I keep seeing articles comparing S&P 500 return versus the IRR of a VC fund, but none seem to compute "IRR" for the S&P 500. That is, they all seem to assume $1 invested at t=0 in S&P 500 (and I assume total return, so reinvested dividends), and then compare that to venture investing.
Except an $100M fund isn't $100M instantly deployed. The investors are putting probably $20M/yr into it via capital calls. That makes a huge difference in IRR.
This isn't to defend the particular investments or performance of any firm, but it does seem like the reporting is quite poor. Even taking the time to compute a "what if each year you invested 1/5th into the S&P 500" would be a marked improvement. But you definitely don't get to say "The 2010 Andreessen Horowitz fund performed slightly better than investments made in the S&P 500 in the same year" (as the article does).
I agree that it's not an apples-to-apples comparison, but it's arguably somewhat in A16z's favor. This is because the fund manager has the ability to defer capital calls as long as possible, because that starts the IRR clock ticking.
This is fairly little known, but large fund managers like A16z have access to "capital call lines of credit," provided by specialty lending arms of banks. These are loans secured by the commitments from highly-creditworthy institutional investors that allow the fund manager to fund expenses and investments by drawing down on the LOC instead of making a capital call. This allows the fund manager to push out the IRR clock even longer, and effectively levers their returns.
A more effective comparison might take into account both your comment regarding deployment period, and also the effect of investing on margin with the CCLOC.
If the fund manager leaves your money sitting around in cash for a few years, that impacts your returns, since you could have put that money to work elsewhere. Including the gradual investment as part of IRR is the correct thing to do.
Sorry for not making it clear: they don’t ask for the money, until it’s “needed”. The term of art is “capital call” and while it’s possible to call at any time (perhaps you want to make a huge investment and you don’t have the cash currently), it’s usually somewhat spread out. The “default” behavior is an even-ish set of calls over say a 5-year period for a (nominally) 10-year fund.
Isn't the money in some sense tied up if it has to be ready for a capital call? At minimum it should be in some relatively low risk liquid investment. So there is opportunity cost regardless of whether the investor or the fund holds it until it is deployed?
Yep, but that’s equally true of any IRR calculation (for any IRR, you should compare to the risk-free rate or alternatively some other equal-risk benchmark). As an example, perhaps the real comparison for VC investment should be to having your “capital to be called” in the S&P 500 while waiting for capital calls. Except, as you allude, that’s quite risky for “you are required to deliver” (and IIUC, often within days). That makes the easiest assumption some sort of money-market fund or treasury something something.
In any case, the $X in S&P 500 at t=0 versus the IRR of a venture fund is not an apples-to-apples comparison. There are many ways to meet your capital calls, and I suspect that sophisticated investors aren’t keeping the cash in their checking account waiting for an email.
I agree it isn't apples to apples. I wonder if it is nevertheless the best comparison, though, because they're two fundamentally different beasts.
If I have a lump of cash I need to deploy, I can buy lump sum SPX at t=0. I can't call a VC and say "here's X money, invest it all immediately."
The fairest comparison probably would be lump sum SPX against a blend of VC IRR and money market, converted from one to the other at a typical capital call rate, but as the number of variables increases so do the number of assumptions, and given the low returns on money market funds I don't see how the extra complexity adds much to the story.
Kind of, it just needs to be sufficiently liquid that you can meet the capital calls as they come. If you've got 10% of a portfolio committed to a VC fund, you can keep the rest in the S&P and just sell shares as needed. However, you would have series of returns risk of the capital calls came during market crashes.
If you've got 50% committed, then you need to do even more careful cashflow planning.
Ok. Picking on a firm because it didn't beat the 500 is harsh. Beating the 500 is damn hard. An 500index fund is basically a collection of large monopolies that extract cash.
Yes, but then why invest in an actively managed fund at all, as an LP? whole premise is that fund managers can beat an index such as the S&P500 or total market etc.
LPs don’t invest in funds to beat the market. By the time they are investing in VC they have already have millions in traditional investments like index funds, real estate, etc. VC investments are a high-risk, high-reward play.
Individual VC investments are a high-risk, high reward play, but for an LP, investments in large baskets of VC investments via VC funds are ideally intended to collect an uncorrelated risk premium. The idea isn't to simply to beat the SP500, its to diversify, and thus improve risk adjusted returns, of their broader investment portfolio.
For large institutional LP's like pension funds, endowments, charities, etc, they need steady smooth returns that they can draw upon year after year to fund their beneficiaries. In particular a down year really hurts them since they will have to draw down on their principal. This is a very different risk calculus to an individual saving for retirement who can stomach 30-40 years of stock market volatility with a good probability of having enough money at the end of their career.
So rather than chucking the bulk of their fund in to the asset with the highest expected returns as an individual might, these institutional investors buy a big basket of very different return streams (i.e. as uncorrelated as possible) to smooth out the bumps in each one.
Shh... if all people do is invest in funds made up of existing blue chips, nothing new will ever be funded again and we'll end up with an economy of nothing but stale old Soviet bureaus.
This article is absurd. It says the 2011 fund return rate is 12%, which isn’t spectacular. That’s the whole point of the article.
Except the 2011 fund includes stakes in Airbnb and Stripe, two massive companies that have yet to go public.
So, sure, at the moment the returns aren’t awe inspiring. But give I can’t imagine anyone at AH is losing sleep over the long-term success of that fund.
Please correct me if wrong, but these are likely measured via the book value of the investments which reflects the private market valuations of the companies you mentioned.
Those are "opportunity funds". "Parallel Funds" normally are funds that invest at the same time and in the same fund percentage as the main fund. These are often setup as a way to group together LPs; for example you may have a main US fund and one in the Caymans that's for foreign investors. They function as a "Fund Family" and because they make the same investments at the same percentages, they normally have the same returns.
Looking at the SEC filing for Pinterest, you can see this is the case for a16z. Fund III and Parallel Fund III invested in the same rounds, always at the same proportion:
"Poor investments can contribute to a fund’s weak performance, but so can decisions to put too little money into startups that ultimately turn into smashes."
I'd guess the problem is the opposite. Too much money into startups that fail.
VC is hard now because there's so much money chasing startups. The valuations they're getting are absurd. When you pay twice as much, your hit gives you have the return and your failure twice the loss. VC has been a pretty poor investment historically, and the high valuations now are really hurting.
Note that this data is from preqin and doesn't include all funds, just those that self report or have LPs who publish returns of funds they invested in
Because the volatility and risk profile of the two investments is not the same.
The index will pay you beta, which is the market return - no more, no less. If you build a portfolio where you combine an investment in the index fund with other, different investments, you can build many different portfolios which all have different risk profiles. This is useful, as people have different needs.
For example, if you're happy for your money to be locked away for a long period of time (say you're running a university endowment fund), you may be better suited to these longer term, illiquid type of in investments.
With VC you are actually investing in new businesses whereas with stocks, unless you are buying at IPO, you are just buying a piece of a pie. And the upside is much greater. What these results show is that a16z maybe needs to be more discerning but not that the business model is unsound
Perhaps, but we don't really know how much risk many of those VC funds are actually taking. With liquid, publicly-traded stocks we can sort of use variability of returns as a proxy for risk. But there's no equivalent good way to really quantify VC fund risk. Sure you can do risk modeling but it's just an educated guess.
Yes, because they failed to use a possessive here. The headline should have been “Andreessen Horowitz’s Returns Slip”. Then we’d parse Returns as a noun rather than a possible verb.
For comparison: Founders Fund has an IRR of ~55%, at ~$1B AUM scale. It tends to invest in fewer companies -- with much higher bar and conviction -- and its portfolio has a much lower failure rate than competing funds (of course failure rate doesn't matter as much for VC returns, but it's still an interesting fact).
I have no connection to FF whatsoever, but have learned a lot from the way they invest and much prefer their model to the spray and pray style (YC, Ron Conway, A16Z, etc).