The falling returns to venture capital

Noah Smith has a very interesting post on this topic, here is one picture, explained more at his link:

Noah wrote:

…that graph sure looks like a structural break to me.  Something looks like it broke the VC business model after the dot-com crash. Maybe the new tech bubble (Facebook, etc.) will pump those returns back up, but there have been some big IPOs and some big acquisitions in Tech Bubble 2.0, and VC returns haven’t really bounced back yet, so I’d be cautious. What’s more, I’m starting to read about a slump in venture funding…Could this be the (temporary) end of the VC industry? If so, is it a harbinger of technological stagnation, or simply the passing of a financial fad?

He is drawing upon this recent paper by Harris, Kaplan, and Jenkinson (pdf).  As I remarked at a party recently: “We are all stagnationists now.”

Addendum: Tim Worstall adds comment.


With the recent JOBS act, you will see VCs offloading some of their properties to the unsuspecting and uninformed. VC activity will go up, and new shareholders purchasing IPOs will eventually assume the VCs bad risk and their shares will go down.

Think CDOs.

Instead of VC activity, what I look for is persons in the industry investing in new startups, not VC investors, as a measure of real world worthiness. In the area where I live, we have two very strong high tech specialties, and the executives and engineers of the successful businesses that started these specialized businesses have since taken their money or retired, and have invested in new start up businesses in the same technological field. I look at what they invest in, and what they don't, and whether they are investing at all.

VC funding and activity is a function of liquidity and extra cash from others. Investment by industry insiders is a function of knowledge. Take your pick.

Here's a link to Esther Dyson on Vox making the same point:

In tech these days, the goal for startups is to get acquired rather than IPO.

We have returned to the old model where once you have innovated and created a market, one of the big guys will step in and crush you unless you are bought first.

So in essence buying Apple et al. is like buying a 90's VC fund, because the path now involves a venture being bought by a large player and that player reaping most of the returns from the venture.

"In tech these days, the goal for startups is to get acquired rather than IPO."
You can put the blame for that squarely on the government. SarbOx is a travesty.

It was actually illegal for most people to buy Apple stock at one point early in their history. I wonder how much of the VC decline is due to the playing field being increasingly level.

I've been responsible for SOx compliance reports. I know the waste associated. But it is the dotcom boom/bust coupled with adventurous tech companies that killed the startup ipo. And bless us all, because most dotcom IPOs were idiotic. Selling technical infrastructure to a company that can handle the business is the better bet for everyone besides us programmers. We love that an IPO that turns into gold but are generally left out of any buyouts.

Before 1998, a $50 million VC fund was a pretty big fund. In the 1998-1999 window, we saw the first $500 million VC funds and the industry became a "style box" for the institutional investors of all stripes. Allocations went from 2% of a portfolio to 5 to 10%. I think a big part of the fall in returns is a function of too much money chasing too few good ideas.

My thoughts too. Initially venture capital was rare and only the better ideas got funded? Is there a way to test the hypothesis; what's the size of the venture capital market then and now?

The question also is: what is venture capital.

If an engineer goes without a salary or takes a low salary in exchange for stock, is that venture capital? I look for "in-kind" venture capital where the in-kind investors have market alternatives for their time..

I more or less agree. How much money got put into VC in the late 90's? Given returns over the preceding years I'd guess a ton.

That means more deals get funded, lowering returns, even if all the new participants know what they're doing. But if there are lots of new VC firms, as well as new investors, some of them aren't going to very good, hurting returns even more. And I bet a lot of marginal operators did come in.

To me the graph shows nothing but the time-tested heuristic that as a industry matures, returns decrease and it gets hard to make super-normal profits any more.

I don't see Venture Capital as any sort of exception to this historical trend.

OK, except for two things:

1. This was pretty sudden, and associated with the end of a related boom.

2. Adjusting for risk it may well be that returns have been not just normal but sub-normal.

So, speaking very broadly you may be right, but I do think the boom very much hastened the drop. I'd like to see what the graph looks like going further back.

I think this is part of it, but there is more on the supply side, namely that a favorite pastime for post-buyout/IPO entrepreneurers has been to create yet another venture capital fund with their newly created wealth.

Ultimately this is why I don't think VC returns will ever recover--its a "fun" business: finding cool new companies, working with smart young people, being constantly on the cutting edge. As long as it doesn't lose tons of money, there will be many wealthy individuals perfectly happy with substandard investment returns.

What is the after tax return, especially if you are forward looking?

well, you have a crash, followed by a legal regiem shift that made going public far too expensive. Imo, ir you repealed SarBox you would see some increase in both the variability and the returns of the funds.

SarbOx, I mean, not SarBox.

What fraction of VC funds are public? I always thought Sarbanes-Oxley burdens were mostly a medium-large corporate issue. Aren't VC-funds and startups typically smaller and leaner where Sarbanes-Oxley is not so taxing?

Yeah, but that doesn't stop someone from saying Sarbane-Oxley is the problem when it doesn't apply. Read the playbook: all efforts to make better disclosure are bad. Here's the proof: you wouldn't invest in this stuff if you knew about the risks. Ergo, there will be less investment. So what. Less wasted resources.

It's not the VC funds, it's their investments. The home runs of venture capital come from taking a company public, with an IPO (essentially turning it into a medium to large company (in which the VC group owns say a 5% stake). The singles and doubles are selling the same company to an already public firm for a much smaller premium. If Sarb-Ox makes the cost of reaching the home run, exceedingly expensive/rare, VC returns will be impacted (they were already a business of 99 strike outs paid for by a single home run).

bluto, Ask the question: If Sarbox disclosure makes the homerun less rare, is that good or bad: if disclosure makes you not want to purchase because the disclosed risk is too high, or the books are cooked, are your better or worse off knowing of the risk or not.

If you think you are better off not knowing the risk, or not knowing the finances, I have a bridge...

While you make some good points, I think the complaint is that the disclosure is expensive and requires revealing strategically significant information that would be better kept private. Neither of those complaints have anything to do with a desire to hide wrongdoing. They justify reluctance to go public in completely legitimate firms.

There's also the point that sarbox is pretty much ineffective in uncovering fraud or promoting fairness, but that's society's problem and doesn't directly discourage IPOs. So it's a sidebar to this conversation.

Finch, One can make the opposite argument as well: That after Enron/MCI/Aderson Consulting, that persons were NOT going to invest in start ups, much less some Fortune 500, because disclosure/regulation was so bad, and that consequently our financial system would be worse off.

As to disclosure, Sarbox doesn't require you to list trade secrets, customers, etc.

Finally, It is interesting that you often hear that the reason US stock markets developed, and the reason foreign firms want to be listed here, is that we have regulation, and that foreign firms would not acquire a premium in their own market for good behaviour, management, etc., because there is a lack of regulation in their market.

The old model was VC got their return when a company IPO'ed. There has been very few tech IPO's since SarbOx. So the new model is design your startup to be acquired by a big tech corp (MSFT, GOOG, CSCO). Those exits are the new IPOs for tech startups.

Also without the tech IPOs, there was no new money being re-invested new startups for a while. That is why SarbOx created a downward spiral.

The other big shifts in VC business model is the lower capital required for the startup of software and web service companies. With cloud services and virtual private hosting services, you don't need a large upfront investment to keep up with demand. Pay as you Grow. So less VC money need, less return when Facebook, Google, or Microsoft buys the successful web startups.

Right because only Doc Merlin and Republican operatives complain about SarbOx, and it clearly fixed the economy and has been a boon for investors.

Why credit any of this to SOx? Facebook, Google, Microsoft, et al. can undercut VCs with real money. VCs expect more and on average pay less.

That graph is bothering me. I agree with fogcity. Finance theory says that marginal returns will decrease as overall investment increases. Global VC capitalization could have easily been tracked in the same graph.

Secondly, the graph starts in the midst of the telco/.com bubble. The interpretation of the graph is "old norm" versus "new norm" (structural break?) when it should be "bubble" versus "post bubble".

Finally, tax law has driven the big returns to leveraged buyouts. These takeovers that don't need high-tech or innovation to be successful -- just the ability to screw over current stakeholders and utilize tax arbitrage. I guess that area is not represented here (in the graph), but it had captured the imagination of the financial sector before our recent troubles.

A few comments:

1) None of the large gains the venture capital industry has made from social media are in the data set yet, since it ends in 2008. Once this are included the numbers will look much "better."

2) The venture business model is a fund with 10-20 investments (most of which lose money) but one "home run" with a 20x+ return; this is also true in aggregate. For example, Google represented a extremely high percentage of the aggregate profits of the ENTIRE venture capital industry in the years around its IPO

3) The same will be true with Facebook, which is expected to have a market capitalization of $100bn, and may represent $20bn of profit to the venture industry. (Not sure if this is exactly the right number, but it's public if someone digs through the S-1). Mind you, the entire amount of capital committed to venture funds is only ~$200bn

4) Maybe some gains that once accrued to venture funds now go to individuals like Peter Thiel and Reid Hoffman?

These numbers include the expected returns to social media companies. For the last couple of years, VCs (and buyout firms) have been required to have their valuations of portfolio companies, which they send out quarterly to their LPs, approved by external sources. The Kaplan data includes these valuations for unrealized investments (i.e., investments that have not gone public or been acquired). So, for instance, the return to their investment in Facebook is included in the returns of Accel's appropriate fund using the March 2011 valuation of Facebook.

Tim, yes you are right that some of the gains may be included, but much of the gains in facebook have taken place in the last 6-12 months, which I'm certain is not captured. (Venture funds normally lag at least one quarter in providing valuations)

Probably 4)

Founders are cashing out at a much later stage than they used to, and that isn't just due to SarbOx; it just costs less nowadays to run a software startup. Once upon a time, even retail investors could get in on the ground floor: if you invested $10,000 in the Microsoft IPO you'd have been a millionaire fifteen years later. But retail investors have long since become superfluous, nobody needs them as a source of funding anymore. The twist is that now venture capital is increasingly finding itself in the same position: they're coming in at a much later stage and a much higher valuation. Angel investors have taken over the original ecological niche, and they're often invited to invest as much for the stamp of credibility and for their advisory role as for the actual funding itself.

So most investors can't beat the S&P 500 index consistently, especially after reflecting fees. The average fund will underperform taking fees into account. Why should venture capital be different?

I view this as vindication for the Efficient Capital Markets Hypothesis.... which probably doesn't need any more vindication.

Long-term 20%+ net returns for VC investors would be vindication of EMH given the riskiness of the asset class.

But finance theory says that riskiness = correlation with the market portfolio, not intrinsic volatility. Not that anyone in the real world thinks this way.

To be precise, finance theory says riskiness = Covariance of an asset with the market over variance of the market (known as "beta"). VC is a very high beta asset class, although it's impossible to measure directly because VC assets aren't publicly valued every day like stocks and bonds.

But Beta doesn't even persist

I'm not sure this is fully understood: we are not looking at a continuous time series, but discrete categories of funds raised in particular years and then invested over time ("vintages"). It's not surprising that 99-01 vintages were terrible and earlier vintages were great because of the dot-com bubble. But any VC investor would recognize that recent vintages could be explained in terms of the "J-curve" effect. VC investments have a long gestation period and are difficult to value in the absence of actual investment events. Recent investments tend to be held close to cost, and their ultimate performance won't be known for years. In the meantime, fees cause negative performance.

Nonetheless, I think there's strong evidence that VC is a terrible asset class outside of a select few funds that consistenly perform well. And there is no doubt that Sarbox made the VC investment environment more difficult.

The best case to the contrary is made here by academics Josh Lerner as well as Kaplan, one of the authors of the paper cited in this post

Paul Graham has written many excellent essays that, taken together, offer some insight into this. His company (Y-Combinator) exists in a middle ground between angel investors and VCs. Having read most of what he's written in the past few years, I'd guess that part of the explanation is: The cost of starting a startup is approaching zero, therefore the need for massive cash infusions early on has gone down in many cases, therefore the VCs have less deals to chase and will tend to drive each others' margins down through competition. By helping startups get from their first few days of existence to a relatively stable state in a short time (and with a relatively modest exchange of cash for equity), Y-Combinator is becoming a massive disrupting force in the startup world.

The Great Stagnation certainly ranks as the biggest thing that I've been convinced to change my mind on in the past couple of years. Ya got me, Cowen...

Say what you will about me, but don't say I'm not open-minded! ;-)

This graph needs to go back further. Why does it start with the tech bubble?

It could also be because the technology has gotten a lot cheaper, there's better market infrastructure and better processes. I'm a big fan of Eric Reis' Lean Startup model, and with coding tools like Ruby on Rails (or the latest version), app stores, and cloud hosting, it gets pretty easy.

Most web or mobile startups these days need FAR less capital to get through the learning phase and start to earn revenue, so the bad ideas can be vetted without VC involvement. A startup doesn't go to the VC's until they have a revenue chart that makes sense and and a clear sense of what money they need. That's a far less risky investment and probably doesn't earn the returns they used to get.

As someone who worked for a consulting company that used to hoover up our client's VC checks at pretty high rates, it's nice to see how flexible the industry is now.

Man I wish I was at that party.

Things like this are why economists (and I am speaking as one) are *so much fun* at parties:

Tyler says: As I remarked at a party recently: “We are all stagnationists now.”

Comments for this post are closed