Marc Andreessen Is Wrong. The IPO Isn't Dying.
Marc Andreessen recently lamented the death of the IPO in the United States, blaming over regulation and short sellers as the reason why the general public no longer gets to share in the spectacular capital returns that earlier technology stocks like Microsoft and Amazon delivered.
While over regulation in the US with regimes like Sarbanes Oxley is a major problem, the data from the National Venture Capital Association shows that over the last five years that both the number of IPOs and the total amount offered is actually in an uptrend. In fact, the first quarter of 2014 saw the strongest three month period for new listings since 2000, and Alibaba is imminently coming to market which will lift the trend in 2014 significantly.
While I agree that regulators in the US have been cutting off the nose to spite the face, I've certainly never heard someone say I'm not going to take my company public because the short sellers are scary. Shorts selling, or borrowing stock from someone else to sell is an incredibly dangerous game. When stocks go down, the lowest they can drop to is zero- but they can go to the moon if a stock suddenly gets hot and sellers decide they no longer want to sell. This is called a short squeeze, and it can drive a short seller bust in an instant- just look up the story of Volkswagen 911 to get a taste. In particular, I can't see why Marc has an issue with Regulation Fair Disclosure, because the shorts will still be spreading falsehoods regardless of whether you're allowed to say anything. Also, since April last year you're allowed to tweet or use social media to debunk rumours, and I don't know why you wouldn't want to tell everyone if it bothered someone so much that they gave you a call to discuss the issue. Reg FD is very similar to the regime in Australia of Continuous Disclosure that has existed for years.
Aside from over regulation, the real reason that stocks appear to be IPOing later to Marc is that he and his top performing peers have been raising bigger and bigger funds, and keeping the spectacular gains that technology companies can deliver all to themselves. This is partly explained by the necessity of having to put more and more money to work and not having many options on how to put all these billions of dollars to work in a way that will generate a return.
VCs make their money from a remuneration structure called "2 and 20"; a management fee (typically around 2% per annum of the funds under management), and a performance fee (typically 20% of the gains the fund makes, over a hurdle rate like 3-5% over a public markets benchmark).
An unprecedented exposé of the venture capital industry was recently published by the Kauffman report, which is probably the most detailed analysis of the venture capital industry, written by an organisation that has probably invested in more venture funds than anyone. The report revealed that venture capital itself is a very tough business unless you are one of the select few funds with a strong brand. It found that venture capital returns haven't significantly outperformed the market since the late 1990s, and since 1997 less cash has been returned to investors than has been invested in VC.
This is, unless you are on of the few top venture brands. These funds benefit from what is known as an information cascade, which is a self-reinforcing loop that because you have a top brand, all the best deals flock to you, and as a result you get the best return, and hence maintain the best brand. As a result, some of the very best funds charge a 30% performance fee, making even higher windfall profits.
While the average venture fund size in the US as a whole hasn't risen, the concentration of funds into the top firms has skyrocketed. Historically, the top 25 venture firms raised 30% of all venture dollars. During the first half of 2012, the top 10 firms raised 69% of all dollars.
Some venture firms have made windfall profits for the partners as a result of this from simply pyramiding the size of successive overlapping funds.
As an example, consider Cowboy Partners raising a $100 million dollars for its first fund. The 2% management fee would equate to $2 million a year, and that would pay for a nice office and indoor putting range, but more importantly dictate how many investment managers the fund could afford. In the case of our $100 million fund, the $2 million would pay for maybe 3-4 investment managers, finance and secretarial staff. These managers only have so much bandwidth, so maybe you'd invest in 10 or so companies, and over the seven year life of the fund you'd put on average $10 million or so into each company.
The cardinal rule of venture investing is the golden rule- he who has the gold rules. This is otherwise known as never run out of dry powder for follow-on investments. If you run out of dry powder, you're going to ski into the rocks because with private, illiquid companies you have no easy way to get out in a hurry and later investors can screw you thanks to changes in control and deal terms like liquidation preference. So if you have $10 million to put into a company then you'd put maybe $2 million in the first round, see how the company goes, maybe $5m in a Series B and hold the rest for later rounds. Good companies you'd put a bit more than average in later, and you'd cut your losers pretty quickly.
What the Kauffman report showed is that venture capitalists have also been making windfalls from the management fee by raising successive, larger funds before profits had been realized in their earlier funds. So, after 18 months of investing out of the $100 million Cowboy Partners I fund from which they made a management fee of $2 million a year, they would raise $500 million for Cowboy Partners II, from which they would make $10 million a year. Eighteen months into that they would raise $1 billion for Cowboy Partners III, from which they would make $20 million a year. So by this stage they are making $32 million a year for doing bugger all. Bear this in mind when considering that the Kaufman report stated that the average VC fund barely manages to return investor capital after all fees are paid.
The good funds like Marc's (his first $300 million fund returned twice the invested capital) can raise bigger and bigger funds from being good operators and benefiting from concentrated deal flow via the information cascade. Others benefit from investing in illiquid private companies for which there is no floating, transparent share price. Eighteen months into Cowboy Partners I, they are out raising for Cowboy Partners II, but at this time there's very little to show in terms of tangible financials for most of the companies their first fund invested in. Many of the companies will have yet to generate any revenue at all. However a lot of these investments will be in the hype stage where they'll be the darlings of TechCrunch and Hacker News and get a lot of press. The metrics used to raise the next fund can be gamed if early in the first fund's life a few quick-flip exits or valuation write-ups occur. This allows 'manipulation' of the Internal Rate of Return (IRR) of the fund. For example, a company that is sold and returns more than twice the invested capital in two years generates a 41% percent IRR, but the same multiple generated by a sale in year ten results is only a 7.2% IRR. Josh Lerner, a professor at Harvard Business School and leading researcher on venture capital said, "When you look at how people report performance, there’s often a lot of gaming taking place in terms of how they manipulate the IRR". IRR together with other funny metrics like 'vintage year' and 'top-quartile' performance is then used as marketing to raise a new, larger fund. These metrics are self-referential and relative measures which compare the fund to other funds, but don't provide any information as to whether the fund is actually hitting the performance hurdle of 3 percent to 5 percent annual returns above the public markets that most investors expect from illiquid, risky venture capital investments.
The problem at this point is you still have limited bandwidth in terms of the deals you can do. Even a rockstar like Marc can only hire a certain number of rockstars to work in his fund. As a result you're investing in a finite number of companies because you still need to do the work to ensure the investment is as successful as it can possibly be. Also there's only a finite number of good companies worth investing in.
The problem becomes what on earth do you do with all this money? As we know in the public markets, the ability of fund managers to generate alpha (returns above market) gets smaller and smaller as their funds under management get bigger and bigger, because the universe of investments you can make gets smaller and smaller.
It's tricky. The Kaufman report showed that over the last twenty years that only four of thirty venture capital funds with committed capital of more than $400 million that they invested in delivered returns better than those available from a publicly traded small cap common stock index. Given the fact that investors in venture funds (called limited partners or LPs) can't generally sell their positions and have to wait many years for the venture fund's positions in companies to be realised (you cut your losers early and double down on your good ones), unless you're investing in a top fund, you're better off just investing in the public markets.
So investing the funds is not easy. In general, you can't give $100 million in a series A to a company, although Marc is giving this a go with deals like the $100 million Series A in Github. You can try to spray and pray into early stage companies like what YCombinator and 500 Startups are doing, but the actual cost of starting Internet and software companies is dropping quickly. Thanks to the combination and interactions of open source, Metcalfe's Law, Nielsen's Law and Moore's Law, you only need to drip feed in $20,000 or so to get a feeling for whether the team can execute and the product or service has legs. As a result, early stage investing has become quite capital efficient; investing in 500 early stage startups at $20,000 a pop only takes $10 million. While this isn't going to scratch the sides of a billion dollar fund, it's certainly going to denial of service the bandwidth of the general partners in the fund to deliver good advice and mentoring to each company.
Getting a return on spray and pray investing however is tough. Returns on portfolio investments are incredibly skewed and approximate to a power law distribution. Your losers will go to zero, with mediocre ones you are lucky to get your money back and great companies return 3-10 times. Peter Thiel describes how the maths works in portfolio of venture investments as "to a first approximation, a VC portfolio will only make money if your best company investment ends up being worth more than your whole fund". If you are investing in 500, or even 100 startups, this can be tricky. As a case in point, Ron Conway, a "super-angel" of Silicon Valley, had Google in his second angel fund and it's not quite clear if investors lost money or were lucky to get it back. Sure, Andreessen Horowitz made $78 million off a $250,000 investment in Instagram, but he would have to find 51 more Instagrams to return the $4 billion in assets under management investors have provided if they solely invested this way. If you're savvy enough you may however be able to cut a few of your losers early on through the growing trend of 'acquihiring' them off to a big corporate and stag your IRR early on (great for your next fund).
Unless you start considering private equity type deals like Marc's well executed $1.9 billion Skype buyout from eBay, the only way to deploy capital from these billion dollar funds is to invest greater and greater amounts into later and later stage companies. This phenomenon has seen the rise over the last number years of gargantuan rounds; $1.5 billion in Facebook at a $50 billion pre-money valuation, $1.2 billion in Uber at a $17 billion pre-money, $950 million in Groupon (of which $345 million was secondary) at a $4.75 billion pre-money and $740 million in Cloudera at a $4.1 billion pre-money.
This has meant companies funded by the top firms in the US have been delaying going public, and it has been helped by the JOBS Act which increased the maximum number of shareholders from 500 to 2,000 that you can have until you're forced to go public.
Of course, a lot of these companies invested in can't actually deploy all these funds effectively, so large percentages of the rounds are not primary (new) issuance but secondary sales where the founders are cashing out large amounts of money before going public. From a founder's perspective this is smart when 99.9% of your wealth is tied up in one asset.
The end result is the general public missing out on the spectacular gains that were experienced in the listed technology companies of yesteryear. Even though eBay's share price went up a spectacular 163% on opening day, if you bought shares on market after this rise and held on until today you'd have made over 3,500%. If you bought Amazon the day after it listed, you'd be up over 27,000%, and if you'd bought Microsoft at IPO in 1986, you'd be up 66,500% today and 3,000% in the first eight years alone. Unfortunately for the general public, not only are these returns being kept by venture capitalists to themselves, but the chance of them being able to invest in a good venture firm is between buckley's and none.
In my next article I'll be writing about a market that I think is about to boom for technology listings, with companies at a size more reminiscent of the US technology IPOs of yesteryear; with market capitalisations in the hundreds of millions instead of the tens of billions. This market is already a top 5 market globally, it's as big as NASDAQ for equity capital raisings. It's where I floated my company, Freelancer.com, in November of 2013. It's called the Australian Securities Exchange.
Matt Barrie is an award winning entrepreneur and the CEO and Chairman of four-time Webby award winning Freelancer.com, the world's largest freelancing marketplace. Prior to founding Freelancer.com, Matt was founder and CEO of Sensory Networks Inc., a vendor of high performance network security processors, which was acquired by Intel Corporation (NASDAQ: INTC) in 2013.